# Pagehive Help: every guide # Getting started ## What Pagehive is https://help.pagehive.space/guide?g=what-is-pagehive ## Live pages, run by your agent Pagehive gives your AI agent a place to build live pages: a dashboard, a tracker, a report or a simple website. The agent keeps the data current, and the page answers back when someone clicks, fills in a form or asks for a change. ## There is no site builder You don't drag boxes around. You tell your agent what you want, and it writes the pages, loads the data and publishes. When something should change, you ask again. ## What a site is made of - Pages: HTML with a little template code, one per address, such as / or /pricing - Tables and values: the data your pages show, kept on Pagehive - Files: images, fonts, PDFs and short videos - Visitor actions and agent commands: the ways a page answers back ## Where things live Every site gets its own address on pagehive.space, such as daily.pagehive.space. You manage your sites, agents and keys in Your Hive at pagehive.io/hive. ## Nothing to host Pagehive serves the pages, stores the data and keeps sites fast. There is nothing to install and no server to run. ## Create your Hive https://help.pagehive.space/guide?g=create-your-hive ## Your Hive is your account A Hive holds your sites, your agents and their keys. You sign in to it at pagehive.io/hive. ## Sign up 1. Open pagehive.io/sign-up. 2. Pick how you sign in: a passkey, Google, GitHub or a code sent to your email. 3. Your Hive opens. It is empty until your agent builds something. > Every Hive starts on the Free plan. You can upgrade to Basic, Pro or Ultra any time from Your Hive. ## Or let your agent start Your agent can sign up on its own and build a first site. It then gives you a claim link, and you sign in through it to keep the site. See Claim a site your agent made. ## What you see in Your Hive - Sites: every site, with its address, visibility and settings - Agents: every agent, its connected apps, keys and level - Settings: your passkeys and other ways to sign in - Usage: how much of your limits you use ## Claim a site your agent made https://help.pagehive.space/guide?g=claim-a-site ## When this happens An agent can sign up for Pagehive by itself and build a site before you have an account. It then hands you a claim link. Only the link, never its key. ## Claim it 1. Open the claim link your agent sent. 2. Sign in with a passkey, Google, GitHub or an email code. 3. Check the sites the page shows, then keep them. If you already have a Hive, sign in with it and the site joins that Hive. ## Don't wait too long Unclaimed sites are deleted after 7 days. Every page of an unclaimed site shows a small bar with the days left. The bar has no link, because the claim link is secret: ask your agent for it. ## After claiming The agent that built the site keeps working on it. On its page in Your Hive, under Agents, you can give it an API key, change its level or limit it to some sites. ## Your first site https://help.pagehive.space/guide?g=your-first-site ## Before you start Connect your agent to Pagehive first. Connect your agent has the steps for Claude, ChatGPT and Grok. ## Four steps 1. Ask your agent for a site. Say what it is for and who will use it, for example: a page that tracks my reading list, with a form to add a book. 2. Let it build. The agent creates the site, writes the pages to a draft and loads the data. 3. Look at the draft. Your agent can open a view link for you, or you open the site from Your Hive. 4. Publish. Ask your agent to publish. The draft goes live as one version. ## It starts private A new site is private: only you, your members and people with a view link can open it. When it is ready, ask your agent to make it public, or switch it in Your Hive. ## Prompts to copy Sites in Your Hive has prompts to copy, for a simple site or a step-by-step walkthrough with your agent. ## How publishing and review work https://help.pagehive.space/guide?g=publishing-and-review ## Every edit goes to the draft When your agent changes a page, a file or a stylesheet, the change lands in the draft. Visitors keep seeing the last published version. ## Publish makes it live Publishing makes every file in the draft live together, as one version. A new site stays offline until its first publish. ## Data is live at once Rows and values are not part of the draft. When your agent changes data, pages that show it update straight away, without a publish. ## Pagehive reviews every publish Until a site passes its first review, search engines are told not to index it. After that it stays indexed while later publishes are reviewed. If a site breaks the rules, Pagehive takes it offline and sends you a message that says what is wrong and what to change. Fix it and publish again: the new version is reviewed, and the site comes back when it passes. # Connect your agent ## Connect Claude https://help.pagehive.space/guide?g=connect-claude ## What you need A Claude account. It works on claude.ai, and in the Claude desktop and mobile apps once it is added. ## Steps 1. On claude.ai, open Customize, then Connectors. 2. Browse the connectors and search for Pagehive, or open claude.ai/directory/connectors/pagehive. 3. Choose Connect. Sign in to Pagehive and approve. 4. In a chat, open the + menu, then Connectors, and check that Pagehive is on. ## Not in the list? Add it as a custom connector: under Customize, then Connectors, choose Add custom connector, name it Pagehive and paste https://pagehive.io/mcp. The Free plan allows one custom connector. ## Team and Enterprise plans An owner may need to allow the connector first, under Organization settings, then Connectors. You then choose Connect next to it under Customize, then Connectors. ## Then The app shows up in Your Hive, under Agents. Ask Claude for a site. > Checked 18 Nov 2026. If the menus moved, see Claude's own help on connectors. ## Connect ChatGPT https://help.pagehive.space/guide?g=connect-chatgpt ## What you need ChatGPT calls a custom connector an app, and adds it in developer mode, on the web. Developer mode is on the Plus, Pro, Business, Enterprise and Edu plans, not on Free. On Business, only an admin or owner can add one. ## Steps 1. On chatgpt.com, open Settings and turn on Developer mode. It sits under Apps, then Advanced settings, or under Security and login, depending on your version. 2. Choose Create app (or the + button) and name it Pagehive. Paste https://pagehive.io/mcp and choose OAuth. 3. Save. Sign in to Pagehive and approve. 4. In a chat, open the + menu, choose Developer mode and select Pagehive. ## Reading but not changing On some plans ChatGPT allows only reading tools in developer mode. If your agent can read your Hive but can't change it, that is the cause, not your agent's level in Pagehive. > Checked 9 Nov 2026. If the menus moved, see OpenAI's own help on apps. ## Connect Grok https://help.pagehive.space/guide?g=connect-grok ## What you need Custom connectors are on grok.com. If you don't see the Custom option, check your Grok plan. ## Steps 1. Open grok.com/connectors (Skills and Connectors in the sidebar). 2. Choose New Connector, then Custom. Name it Pagehive and paste https://pagehive.io/mcp. 3. Choose Add Connector. Sign in to Pagehive and approve. ## Business and Enterprise On Grok Business or Enterprise, a team admin may need to add the connector first. > Checked 9 Nov 2026. If the menus moved, see xAI's own help on connectors. ## Connect Cursor https://help.pagehive.space/guide?g=connect-cursor ## What you need Cursor on your computer. ## Steps 1. Open cursor.directory/plugins/pagehive in your browser and choose Add to Cursor. 2. Cursor opens and asks to install Pagehive. Accept. 3. The first time Cursor uses Pagehive, sign in to Pagehive and approve. ## Or add it by hand Add this to Cursor's mcp.json: ```json { "mcpServers": { "pagehive": { "url": "https://pagehive.io/mcp" } } } ``` ## Then The app shows up in Your Hive, under Agents. Ask Cursor for a site. > Checked 18 Nov 2026. If the menus moved, see Cursor's own help on MCP. ## Claude Code and API keys https://help.pagehive.space/guide?g=connect-with-an-api-key ## When to use a key Claude Code, Claude Desktop with a config file, scripts and scheduled routines send an API key instead of signing in. ## Create a key 1. In Your Hive, open Agents. 2. Choose Add agent and pick API key, or open an agent you already have. 3. Create the key. It is shown once, so copy it straight into your tool. ## Claude Code Add Pagehive with your key in place of the placeholder: claude mcp add --transport http pagehive https://pagehive.io/mcp --header "Authorization: Bearer bp_live_..." ## Keep it secret A key can change your Hive. Never paste it into a page, a chat with someone else or a public file. Pagehive refuses to save a page that contains a live key. If a key leaks, rotate it on its agent's page. ## More than one connector in one app https://help.pagehive.space/guide?g=more-than-one-connector ## Why you would You may want two agents in the same app: one for your whole Hive and one limited to a single site, for example. ## How Add /mcp/ and a name of your choice to the address for each extra connector, such as https://pagehive.io/mcp/brain. The name is lowercase letters, digits and dashes, up to 40 characters. Every address reaches the same Pagehive. When you sign in, the page asks which agent the connector works as: one you already have, or a new one. ## Agent levels https://help.pagehive.space/guide?g=agent-levels ## Every agent has a level An agent's keys and connected apps do only what its level allows. You pick the level on the agent's page in Your Hive, under Agents. ## The levels - None: it can only check who it is, read the guide and contact support - Read: it sees your sites, pages, files and data - Edit: it also changes rows and runs visitor actions and commands - Build: it also creates sites and changes pages, files and tables - Publish: it also publishes, rolls back, makes sites public or takes them offline - Full: everything, including deleting and sharing - Custom: pick each permission yourself Edit, Build and Publish never delete. Deleting is on only at Full, or in Custom. ## New agents start at None Your Hive's first agent starts at Full so it can build right away. Every agent you add later starts at None. If an agent says it has no rights yet, raise its level on its page. ## Reach and how it works On the same page you choose its reach (the whole Hive, every site except some, or only some sites) and whether it works as you or as itself. A change applies to its next call. ## Disconnect an app https://help.pagehive.space/guide?g=disconnect-an-app ## Where Every connected app shows on its agent's page in Your Hive, under Agents. ## What disconnecting does - The app stops on its next request - View links that app created stop working - The agent stays, with its other apps and keys To stop the agent altogether, remove it on the same page. To pause it without removing anything, set its level to None. # Sites and pages ## Drafts, versions and rollback https://help.pagehive.space/guide?g=drafts-and-versions ## The draft Every change your agent makes to pages, partials, stylesheets, scripts and files goes into the draft. Nothing changes for visitors until you publish. ## Versions Every publish stores a version. Your agent can list versions, compare two of them and show you what changed. ## Rollback If a publish went wrong, ask your agent to roll the site back to an earlier version. That version goes live again at once. ## What versions don't cover - Data: rows and values change live and are not rolled back - Visibility and page access rules: they apply at once and stay as they are ## Pages and partials https://help.pagehive.space/guide?g=pages-and-partials ## Pages A page is one address on your site, such as / or /blog/launch. It is HTML with a little template code that prints your data. Paths can nest, and each path is its own page. ## Partials A partial is a named piece every page can include, such as a header, a footer or a menu. Change the partial once and every page that includes it changes with the next publish. ## Styles and scripts Each site has one global stylesheet, and each page can add its own. Scripts live in their own files under /scripts. The HTML is the page: a script only adds to it. ## Folders or separate sites Use folders, such as /shop and /blog, when sections should feel like one website: they share partials, styles and scripts. Use separate sites when a section should look different or grow large. Every site has its own address and page limit. ## Images, fonts, PDFs and video https://help.pagehive.space/guide?g=files-and-media ## What you can upload Images (png, jpg, gif, webp, avif, svg, ico), PDFs, fonts (woff2, woff, ttf, otf) and short videos (mp4, webm). After a publish, a file is served at /files/ and its name, such as /files/logo.png. ## How your agent uploads - A file on its own computer: it asks for an upload link and sends the file there - A file already online: it gives Pagehive the link and Pagehive fetches it - Many files at once: up to 20 in one go, from its computer or from links. They land in the draft together, as one change Uploading the same name again replaces the file. Visitors see new files after your agent publishes. ## Fonts Host fonts on your site rather than loading them from another service. Pages load faster and visitors aren't tracked by a third party. ## Size Files count toward your Hive's storage. Keep images sized for the screen they are shown on. ## Feeds, JSON and text pages https://help.pagehive.space/guide?g=feeds-and-text-pages ## A page in another format A page whose address ends in .rss, .atom, .xml, .json, .txt, .md, .csv or .ics is served in that format instead of HTML. It works like any other page: it can read your tables and include partials. ## What people use it for - /feed.rss: a news feed readers can follow - /api/items.json: your rows as JSON for another tool - /events.ics: a calendar people can subscribe to - /llms.txt: a plain-text summary for AI agents ## Good to know These pages count toward your page limit. robots.txt and sitemap.xml are separate files your agent publishes, not pages. ## Link previews and site icon https://help.pagehive.space/guide?g=link-previews-and-icon ## Link previews When someone shares a link, X, LinkedIn, Slack, WhatsApp and Discord show a card with a title, a description and an image. Your agent sets these per page. - Title: the page's own title - Description: one or two sentences about the page - Image: 1200 by 630 pixels, PNG or JPG. Previews don't show SVG A page without an image of its own shows the Pagehive card. ## Site icon Without an icon of its own, a site shows the Pagehive hive. To use your own, your agent saves an SVG icon at /favicon.svg and publishes. ## Check it The SEO scan lists pages with a missing title, description or share image. See Run an SEO scan. ## Design mode https://help.pagehive.space/guide?g=design-mode ## What it is Design mode shows your site's draft. You click the part you want changed and write what should change. Your own agent makes the change, and that part updates on your screen. ## Who can use it - The site owner, on Pro or Ultra - Members with the Build or Publish role, if the owner switched on agents for them. Their requests go to their own agent ## Make a change 1. Open your site while signed in and choose the Design tab on the right edge. 2. Click the part you want changed. Pick another page from the page list. 3. Write what should change and choose Send to my agent. 4. Your agent changes the draft. The part updates in place, with a short note of what it did. While a request is open, nobody else can send one for the same part. You can cancel yours until your agent picks it up. ## Pages with buttons, menus or slides Switch the panel from Select to Interact to use the page as a visitor would. Switch back to pick again. Alt-click passes one click through to the page, and Esc takes you back to Select. ## Live or queued - With a Webhook, your agent starts at once. See Connect a Webhook - Without one, the button says Add to queue, and requests wait for your agent's next run ## Publish Changes stay in the draft until someone publishes. With a Webhook, the owner can publish from the panel, and so can members whose agent may publish. # Data ## Tables and columns https://help.pagehive.space/guide?g=tables-and-columns ## A table holds rows Your agent creates tables for the data a site shows: tasks, books, sales, members. Each column has a type: text, number, date, date and time, select, checkbox, a link to another table, or a file. ## Site or account - A site table belongs to one site. Use it for that site's data, its visitor actions and its members. - An account table is read by every site in your Hive. Use it for data several sites show. ## Who can read it - Private, the default: only pages that print it show it - Public: page scripts and anyone with the address can read the rows - Follows the site: everyone who can open the site reads every row - Members' own rows: each member sees only the rows they added Private keeps a table out of page scripts, not out of pages. A page that prints a private table shows those rows to everyone who opens it. ## Load rows https://help.pagehive.space/guide?g=load-rows ## In bulk Your agent loads up to 500 rows in one call. It matches each row on a key column, such as an email or a slug, so a row that already exists is updated rather than added twice. ## All or nothing If one row is wrong, nothing is written and the error says which row and which field. Your agent fixes it and sends the batch again. ## Live at once Rows don't need a publish. Pages that show the table update straight away. ## Limits A table holds up to 10,000 rows, and a row up to 32 KB. Your Hive's total is on the Usage overview in Your Hive. ## Show live data on a page https://help.pagehive.space/guide?g=show-live-data ## On the server A data header at the top of a page names the table, the filter, the sort and how many rows. Pagehive loads those rows before it sends the page, so visitors see the data at once and search engines can read it. ## In the browser A small Pagehive script refreshes rows without reloading the page, for tables that are public or follow the site. Use it for live counters, filters and searches. ## Use both The usual pattern: rows on the first paint from the data header, then a refresh in the browser when something changes. ## Links between pages A page can read a value from its address, such as /project?p=12, and filter its rows with it. One page then shows any project. ## Values https://help.pagehive.space/guide?g=values ## A value is one fact A value holds one thing: today's total, an opening time, a status, a setting. Pages print it, and your agent updates it when it changes. ## Who can read it - Public: pages and page scripts can read it - Private, the default: pages can print it, scripts can't fetch it - Secret: for tokens, API keys and passwords. Only your agent reads it, and no page can print it Private is not secret: anything a page prints, every visitor of that page sees. Store keys and tokens as secret values. ## Charts https://help.pagehive.space/guide?g=charts ## No built-in charts Pagehive has no chart tag. Your agent brings the library it likes, such as ECharts, Chart.js or D3, as a script file on your site or from a public CDN. ## How it fits together 1. The page has an empty element with an id where the chart goes. 2. A script reads the rows, from the page or with the Pagehive browser script. 3. The library draws the chart in that element. ## Keep it on brand Ask your agent to use your colours and to label the data. One chart per question reads better than one chart that does everything. ## Search, filters and paging https://help.pagehive.space/guide?g=search-filters-paging ## Search Columns marked searchable can be searched by words. A search box on a page sends the words in the address, and the page shows the matching rows. ## Filters and sort A page can filter on any column and sort on columns marked sortable. A filter form that sends values in the address works without any script. ## Paging Show a page of rows at a time and a link to the next page. Pages stay fast as the table grows. ## Keep pages light Every page view has a budget. A page that loops over thousands of rows, or loops inside a loop, can go over it as the table grows. Show fewer rows per page and let visitors page or search. # Visitors and members ## What visitor actions are https://help.pagehive.space/guide?g=visitor-actions ## A page that answers back A visitor action lets someone change your data from a page: tick a task, move a card, add a booking. Visitors never write to a table directly. Your agent defines a named action, and the page calls it. ## Each action is narrow An action says exactly what it may change: one checkbox, one column with a fixed set of choices, or a list of fields with limits on length and range. Anything else is refused. ## Who may use it Every action has its own access: - Off: nobody, the default - Owner: only you - View link: you and people with a view link - Members: members of the site with role edit - Anyone: everyone who can open the page Actions that edit, add or delete rows are never open to anyone. Only you, view links or members. ## Safe by design Each action is checked on the server, double clicks never add a row twice, and an edit of a row someone else just changed is refused rather than overwritten. ## Tick, edit, add and delete https://help.pagehive.space/guide?g=tick-edit-add-delete ## The kinds - Toggle: flips one checkbox, such as done - Set: picks one of two or more choices, such as open, doing, done - Fixed: writes fixed values, such as an Archive button - Edit: changes the fields you list, with limits - Add: adds a row from a form - Delete: removes a row. There is no undo ## Extra writes A set action can write more than one column. Marking a task done can also set done on to today, and reopening it can clear that date. ## Dates Today and now follow your site's time zone. Ask your agent to set it, such as Europe/Zurich. ## Your agent hears about it When you connect a Webhook, each action can also reach your agent, so it can react: send a reminder, update a summary, tidy the data. ## Site members and roles https://help.pagehive.space/guide?g=site-members ## Invite by email You invite people to a site, public or private, by email. They sign in with a code sent to that address. Members also find the site under Shared with you in their own Hive. ## Roles - View: they see the site - Edit: they also run actions and commands open to members - Build: their own agent may also change the draft and the site's tables - Publish: their agent may also publish A member's own agent works on the site only when you switch that on for them. ## Their own rows A row a member adds belongs to them. A table can keep each member's rows to that member, or keep shared rows next to private ones. ## On a public site Strangers see the pages as visitors. Give members a small Member sign-in link so they get their own view. ## Staying signed in and signing out A member stays signed in for 30 days after their last visit, so people who keep coming back aren't asked again. On a public or shared computer they stay in only until the browser closes, and at most 12 hours. Your agent can add a Sign out button to any page. It signs that person out of this one site, and works for you as the owner too: you stay signed in to your Hive. ## Removing someone Removing a member ends their access at once. ## Open a site with a PIN https://help.pagehive.space/guide?g=open-with-a-pin ## When to use a PIN A PIN opens a private site to people without an account or an email invite, such as grandparents or a sports team. ## How it works - A site can have up to 10 PINs, each 6 to 12 digits - Each PIN has a role, view or edit, and a label such as Grandparents - Typing the PIN keeps that device in for 30 days after its last visit - On a shared computer, visitors tick the box and stay in only until the browser closes ## Changing a PIN A new PIN signs out everyone who used the old one. Removing a PIN ends its visits and no other. PINs are stored hashed, so write yours down when you set it. ## Good PINs Pagehive refuses runs like 123456 and one digit repeated. Wrong guesses are limited. ## Page access https://help.pagehive.space/guide?g=page-access ## Narrower, never wider A page is as open as its site, unless a page access rule keeps it stricter. A rule never opens a private site to strangers. ## Who a rule lets in - Link: people with a view link, members and you - Members: any member, and you - Editors: members with role edit or higher, and you - Owner: only you A rule can cover one page, such as /notes, or everything under a path, such as /admin/*. ## What visitors see On a members or editors page, a visitor gets a This page is private page with a member sign-in. On an owner page, they get a plain not found, so nobody learns it exists. ## Pages, not data Rules protect pages, scripts and files. A public table is still public, whatever page shows it. Keep members-only data in a private or members' own table. ## Member uploads https://help.pagehive.space/guide?g=member-uploads ## Who can upload You and members with role edit can upload images and short videos from a page: png, jpg, gif, webp, avif, mp4 and webm. View links and strangers can't. ## How it is set up Your agent gives a table a file column and adds it to an add or edit action. The page's form then shows a file picker. ## Who sees the files A file is shown only to people who may read its row. Removing a member, a row or the column removes access at once. ## Limits A file is at most 25 MB and counts toward your storage. # Sharing ## Make a site private or public https://help.pagehive.space/guide?g=private-or-public ## Private A new site is private. Only you, your members, PINs and people with a view link can open it. Strangers see a This site is not public page. ## Public A public site is open to anyone with the address, and search engines can find it once Pagehive's first review has passed. ## Change it Ask your agent, or switch it on the site's settings in Your Hive. The change applies at once. ## Before you go public - Run a security scan and fix anything high or medium - Run an SEO scan for titles, descriptions and share images - Check which tables are public and which actions anyone can run ## View links https://help.pagehive.space/guide?g=view-links ## What a view link does A view link opens a private site for whoever has it, without an account. It opens the pages, scripts and files, and the data the site lets in. ## Getting one The first publish of a private site returns a view link. Your agent can make more. A link is shown once, so keep it somewhere safe. ## Expiry A new link works for 30 days. Your agent can set another date, or a link that never expires. The security scan flags links that never expire. ## Revoking Your agent can revoke a link at any time. Disconnecting the app that made a link also revokes it. ## Lock a site private https://help.pagehive.space/guide?g=lock-a-site-private ## For sites that must stay private In Your Hive you can lock a site private. While it is locked, no agent and no key can make it public, not even while publishing. ## Who can lock it Only you, signed in to Your Hive. Agents can't set or clear the lock. ## Good for Family organisers, internal dashboards, anything with personal data. ## Take a site offline https://help.pagehive.space/guide?g=take-a-site-offline ## What offline does An offline site hides its pages, its data and its view links. Nothing is deleted: bring it back online and everything works again. ## Who can do it You, or an agent at level Publish or Full. ## Limits still count An offline site and its pages still count toward your limits. To free the slots, delete the site. ## Redirects https://help.pagehive.space/guide?g=redirects ## When you need them You moved a site to Pagehive, or renamed pages. Redirects send old addresses to the new ones, so links and search results keep working. ## How they look One rule per line: the old path, the new path and an optional status. - /index.html to / sends the old home page to the new one - /post/:slug to /news/:slug keeps every old post address working - /old-campaign to a full https address sends people to another site ## Good to know Redirects go live with the next publish, and a published page at the same address always wins. A site holds up to 100 rules. ## Run an SEO scan https://help.pagehive.space/guide?g=seo-scan ## What it checks The SEO scan reads each page the way visitors get it and lists what to improve: - Search: whether search engines can see the site, titles, descriptions, sitemap - Sharing: share images, titles and descriptions for link previews - Structure: headings, image descriptions, links to nothing - Weight: heavy pages and images ## Run it before going public A private site is checked as if it were public, so it is ready the day you switch. Your agent can also check the draft before a publish. ## Where to see it Ask your agent to run it, or open the site's SEO tab in Your Hive. Fix high and medium findings first. ## Use your own domain https://help.pagehive.space/guide?g=custom-domains ## Which plans Basic includes 1 custom domain, Pro 10 and Ultra 30. Free has none. ## Connect one 1. In Your Hive, open the site's Domain tab and enter the domain, such as www.example.com. Or ask your agent to connect it. 2. Pagehive shows the DNS records to add: a TXT record that proves the domain is yours, and a CNAME, or A and AAAA records for a domain without www. 3. Add them at your DNS host, then choose Check DNS. Changes can take a few minutes. Once the check passes, the site is served on your domain right away, and its pagehive.space address redirects there. ## example.com and www.example.com They are two separate domains. The first one you add becomes the main address, and the others redirect to it. You can make another one the main address later. ## On Cloudflare Set the records to DNS only, not proxied. Every address the domain points to must be Pagehive's. ## If it stops working Pagehive checks every domain once a day. If one no longer points here, the site goes back to its pagehive.space address until the records are fixed, and your agent gets a message about it. A domain that stays broken for 14 days is released. ## Turn off or disconnect Turning a domain off keeps it with its settings and frees its place on your plan, so you can move to a smaller plan without losing it. Disconnect removes it. # Your Hive and security ## Agents and API keys https://help.pagehive.space/guide?g=agents-and-api-keys ## Agents decide, keys follow Every API key and connected app belongs to an agent. The agent's settings decide what a call may do, not the key. Change the agent, and all its keys and apps follow on their next call. ## Three settings per agent - Level: what it may do. See Agent levels - Reach: the whole Hive, every site except some, or only some sites - How it works: as you, or as itself on the sites it reaches ## Kinds of key - API key: belongs to an agent and does what the agent may - Read-only key: only reads - Site key: an agent that signed up by itself gets one, limited to its own site ## Only you manage keys You create, rotate and revoke keys on the agent's page in Your Hive. Agents never create or change keys. A new key is shown once and stored only as a hash. ## Rotate or revoke a key https://help.pagehive.space/guide?g=rotate-or-revoke-a-key ## Rotate Rotating gives the key a new secret and stops the old one. Use it when a key may have leaked. The new secret is shown once: paste it into the tool that uses it straight away. The agent keeps its level. ## Revoke Revoking stops a key for good. Use it for a key nothing needs any more. ## Where Both are on the agent's page in Your Hive, under Agents. Site keys are listed there too. ## Spot unused keys The security scan flags site keys unused for 30 days. Revoke what you don't use. ## Run a security scan https://help.pagehive.space/guide?g=security-scan ## What it shows The security scan lists every page, table, value, file, action and command of a site, and who can reach each one: strangers, view links, members and members' agents. ## Findings - High: something private that strangers can see, such as a public page that prints a private table - Medium: a private value on a public page, or an action anyone can run - Low: view links that never expire, unused keys, pages nothing links to - Info: the site's settings, links and members ## Settings, not content The scan reads settings and structure only, never the rows or what your pages say. It does not look for keys in pages, so keep tokens in secret values. ## When to run it Before you make a site public, after you add data or pages, and whenever you wonder whether something is safe. The latest report is on the site's Security tab in Your Hive. ## Limits https://help.pagehive.space/guide?g=limits ## Free - 3 sites and 30 pages in a Hive, 10 per site - 2 agents and 2 API keys - 100 MB storage, 10 MB per file - No custom domains ## Basic - 3 sites and 100 pages in a Hive, 50 per site - 2 agents and 5 API keys - 250 MB storage, 10 MB per file - 1 custom domain ## Pro - 10 sites and 500 pages in a Hive, 100 per site - 6 agents and 15 API keys - 1 GB storage, 25 MB per file - 10 custom domains ## Ultra - 30 sites and 2,500 pages in a Hive, 250 per site - 20 agents and 30 API keys - 10 GB storage, 100 MB per file - 30 custom domains See every limit and price on the pricing page at pagehive.io/pricing. ## What counts Only pages count toward the page limits. Partials, stylesheets and scripts are files. An offline site still counts: delete it to free its slots. ## Watch your usage Your Hive's overview shows what you use. Your agent sees a warning from 80% of a limit. ## When you hit one See You hit a limit. ## Delete a site https://help.pagehive.space/guide?g=delete-a-site ## Only you Agents can't delete sites, whatever their level. You delete a site in Your Hive, in the site's settings, by typing its subdomain. ## What happens - The site, its pages and its files are gone - Its slots free up straight away - The subdomain is kept for you for 30 days, so nobody else takes it ## Not sure Take the site offline instead. Nothing is lost, and you can bring it back. ## Delete your account https://help.pagehive.space/guide?g=delete-your-account ## Only in Your Hive You delete your account yourself, signed in at pagehive.io/hive, under Settings, then Delete account. No agent, key or connected app can do it. ## Before you do - Save anything you want to keep: ask your agent to export your tables - Tell members of shared sites that they will lose access ## What happens Deleting your account deletes all its sites and all their data: pages, tables, files, keys and connected apps. Your sites go offline at once and you are signed out. This can never be undone. # Agent commands ## What agent commands are https://help.pagehive.space/guide?g=what-agent-commands-are ## A button that asks your agent A command is a named instruction a page sends to your agent, such as Update the daily overview, or Add to the todo list with the text someone typed. ## How it works 1. Your agent defines the command and puts a button or a form on the page. 2. Someone presses it. Pagehive queues the command. 3. Your agent picks it up, does the work and marks it done, with a short note. The page can show progress: sent, in progress, done. ## Who may send one Like actions, every command has its own access: off by default, then only you, view links or members with role edit. ## Picking commands up Without a Webhook, your agent sees waiting commands the next time it works on the site and asks you before it runs them. With a Webhook, Pagehive starts your agent as soon as a command comes in. See Connect a Webhook. ## Connect a Webhook https://help.pagehive.space/guide?g=connect-a-webhook ## One Webhook per site Each site has one Webhook, set on the site's Webhook tab in Your Hive. Only you set it. Your agent never sees the address or the token. ## Set it up 1. Add the agent to the site and connect it with Pagehive, so it can pick up commands. 2. Create a trigger in your agent: a Claude Code routine with an API trigger, a ChatGPT workspace agent trigger, a Grok Bot routine webhook, or any https webhook such as n8n or Zapier. 3. Paste its address and token on the Webhook tab. 4. Press Test connection. Connected means Pagehive started your agent and it finished the test. ## What it sends Pick which events go to the agent: commands, and optionally visitor actions and form submissions. Generic webhooks are signed, so your server can check each delivery came from Pagehive. ## Messages from Pagehive https://help.pagehive.space/guide?g=messages-from-pagehive ## What they are Pagehive sends your Hive messages: notices about a site, such as one taken offline and what to change, and updates about Pagehive itself. ## Where to read them In Your Hive, under Messages. Your agent sees the unread count too and can read them for you. ## Read together Read state is shared: once you or one of your agents reads a message, it is read for all. You can't reply to a message. For a question about one, open a support ticket. # Troubleshooting ## Common errors https://help.pagehive.space/guide?g=common-errors ## Your agent has no rights The agent's level is too low for what it tried. Raise its level on its page in Your Hive, under Agents. ## The table or value is private A page script tried to read a private table or value. Make it public, or let the page print it on the server. ## The site is locked private An agent tried to make a locked site public. Unlock it in Your Hive first, if you mean to. ## The page can't be shown right now The page went over its render budget, usually by looping over too many rows. Ask your agent to show fewer rows per page. ## Too many requests Your agent made too many calls in a short time. It waits the time Pagehive gives and tries again. You don't need to do anything. ## The day's API calls are used up Each plan allows a number of API calls a day, and the count starts again at midnight UTC. Waiting a minute won't help. Let your agent finish tomorrow, or move to a bigger plan under Billing in Your Hive. ## Something else If an error looks like a Pagehive bug, or doesn't say what to change, open a support ticket. ## You hit a limit https://help.pagehive.space/guide?g=you-hit-a-limit ## What happened Your agent tried to add something past your plan's limit, such as a 4th site or an 11th page on a site on Free. Nothing was written. ## Make room - Delete sites you no longer use. Taking a site offline doesn't free its slots - Merge small pages, or move sections into a separate site - Clear old rows from big tables ## Need more Move to a bigger plan under Billing in Your Hive. Limits shows what each plan holds. If even Ultra isn't enough, open a support ticket with what you are building and which limit you hit. ## Your site is under review https://help.pagehive.space/guide?g=site-under-review ## Not in search yet Until a site passes Pagehive's first review, search engines are told not to index it. Visitors can still open it. After the first review, it stays indexed. ## Taken offline When a site breaks the rules, Pagehive takes it offline. Visitors see an unavailable page, and you get a message saying what is wrong and what to change. ## Bring it back 1. Read the message in Your Hive, under Messages. 2. Fix what it names. 3. Publish again. The new version is reviewed, and the site comes back when it passes. If the decision looks wrong, open a support ticket. ## Open a support ticket https://help.pagehive.space/guide?g=open-a-ticket ## Two ways - In Your Hive, under Support - Ask your agent to open one. Every agent can, even one at level None ## What to include What you did, what you expected and what happened. For a bug, the error your agent saw. Never include keys, tokens, passwords or a visitor's personal data. ## Kinds Bug, feedback, feature request, question, something about your account, or abuse of the rules by a site. ## Answers Each ticket is a conversation. You see replies in Your Hive, and your agent sees them the next time it starts work. # Tools reference Every tool an agent can call, sorted by topic, with the plans that include it. https://help.pagehive.space/tools ## Tools: Getting started ### whoami https://help.pagehive.space/tool?t=whoami Who this key is and what it can reach. bot is the bot behind it; rename_bot renames it. credential is account_key, site_key, or oauth. scope is full, read_only, site, sites, or sites_read; account_wide is true for an account key. sites lists the live sites this key can open. leaves_out_sites true means the agent has every site except some: later sites join, and it can create sites, but it can't manage keys, apps or agents, share or move data, or read account data. support_unread and messages_unread count unread support answers and Pagehive messages; list_my_tickets and list_messages with unread true return them. permissions lists what the agent's level turns off (read, data, draft, delete, publish, share, run, define); a call that needs one fails with insufficient_scope, reason permission_off. At level None only whoami, the guide, support and messages work; every agent after the hive's first starts there, and only the owner raises it, in /hive under Agents. shared_sites lists sites shared with this account where this agent is picked, with role, member_id and allowed_tools, the only tools that work there (others fail with reason member_role; an agent not picked gets agent_not_picked). agent has id, name, description, works_as (you or itself), level, reach, and at level None a note. An agent that works as itself also lists memberships (site_id, member_id, role, allowed_tools) and works on each site as that member. Managing agents and keys is owner_only; sharing data needs an agent that works as you on the whole hive (share_whole_hive). Plans: All plans Parameters: none Example arguments: {} ### get_account https://help.pagehive.space/tool?t=get_account Read this API key's account, tier, effective limits, rights, and usage. A null limit is unlimited. quotas shows each capped quota with used, limit, remaining and warning (true from warning_percent), plus resets_at (the next UTC midnight) on the daily ones; api_calls_per_minute is this key's last 60 seconds. key says whether this is an account key, a site key, or a connected app, and which sites it can open. For a claimed site key, whoami describes the key. list_tiers lists the tiers to compare. Plans: All plans Parameters: none Example arguments: {} ### list_tiers https://help.pagehive.space/tool?t=list_tiers List the tiers this account can see, with their rights and limits: the public tiers and the account's own. A null limit is unlimited. warning_percent is a usage warning, not a hard gate. The account's own tier is on get_account. There is no tool to change a tier. Plans: All plans Parameters: none Example arguments: {} ### request_owner_sign_in_link https://help.pagehive.space/tool?t=request_owner_sign_in_link Create a one-time owner sign-in link that lasts 15 minutes by default (sign_in_link_lifetime_minutes). Requires a full key on a claimed account whose owner has no way to sign in yet (an account claimed before passkeys). Once the owner has a passkey, Google, GitHub, or a recovery code, it is refused with reason owner_signs_in: that owner signs in at /sign-in. Give the owner the link, not the API key. Confirming it starts a session on /hive, where the owner can add a passkey or move the account onto one. Plans: All plans Parameters: idempotency_key | string | | Any unique string (1-200 printable ASCII). A retry with the same key and arguments returns the first call's result instead of running again. Same as the Idempotency-Key header. Example arguments: {} ### rename_bot https://help.pagehive.space/tool?t=rename_bot Rename the agent behind this key or connected app (leave bot_id out, or pass its own id from whoami). Works at every level, None included, and for an agent that works as itself. Only the owner renames other agents, in /hive under Agents (reason owner_only). The new name shows on every site the agent is on. Rows (_created_by, _updated_by), the activity log (actor_name) and the action log show the new name at once, old entries too. Plans: All plans Parameters: bot_id | string | | Leave it out to rename this key's own agent. Only its own id is accepted; the owner renames other agents in /hive. name | string | required | The new name. It shows on every site the agent works on and on the rows it wrote. Example arguments: { "name": "Research bot" } ### get_guide https://help.pagehive.space/tool?t=get_guide The Pagehive bot guide, one topic at a time. Without topic it returns the list of topics with their sizes. topic "tools" lists the tool groups this key can call. Other topics include sign-up (keep the API key to yourself, hand only claim_url to the owner), build-a-site (draft then publish), site-structure-and-limits, values, create-a-table-and-load-rows, query-options, agent-commands, site-templates, support, errors, and worked examples. topic "all" returns the whole guide, which is large. The whole guide is also GET /docs/bot-guide.md and the resource pagehive://docs/bot-guide.md. Plans: All plans Parameters: topic | string | | A topic slug from the index. Leave it out for the list of topics; "tools" lists the tools this key can call, "all" returns the whole guide. Example arguments: { "topic": "build-a-site" } ## Tools: Sites and pages ### list_sites https://help.pagehive.space/tool?t=list_sites List sites on this account. shared_sites lists sites other people shared with this account and let its agents work on, with role, member_id and allowed_tools. Plans: All plans Parameters: none Example arguments: {} ### get_site https://help.pagehive.space/tool?t=get_site Get one site and the files in its draft. current_version_id is the version visitors see. published_at is the first publish time and does not change. last_published_at updates on every publish. The rename object says how many renames are left and what happens to the old address. Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. Example arguments: { "site_id": "01a1…" } ### get_site_stats https://help.pagehive.space/tool?t=get_site_stats Visitor analytics of one site, in UTC days: page views, visitors, visits, bounce rate, average visit length, visitor action events and bot hits (AI crawlers and agents named, kept apart from people), with the previous period of the same length, an hourly or daily chart, and top values for pages, entry and exit pages, referrers, UTM tags, countries, devices, browsers, OS, events and bots. period is today, yesterday, 7d, 30d (default), 90d, 6mo, 12mo or all; or give from and to as YYYY-MM-DD. realtime is the last 5 minutes. access says what the tier allows; a breakdown it does not allow is left out. The owner's own visits are not counted. Today is rolled up every 10 minutes. Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. period | one of: today, yesterday, 7d, 30d, 90d, 6mo, 12mo, all | | The time range, counted in UTC days. Default 30d. Leave it out when you pass from and to. from | string | | First day of a custom range, as YYYY-MM-DD. Use it with to. to | string | | Last day of a custom range, as YYYY-MM-DD. Use it with from. limit | integer | | How many top values to return for each breakdown, such as pages or referrers. Default 10. Example arguments: { "site_id": "01a1…", "period": "7d" } ### create_site https://help.pagehive.space/tool?t=create_site Create a site. The key issued at sign-up can create one site and is then limited to it. A site key that already belongs to a site cannot create another. Creating a site needs an agent that reaches the whole hive (or every site except some) at a level with draft on. New sites are private until update_site or publish_site sets visibility public, unless the owner locked them. publish_site takes optional visibility private or public; omit it and the site stays as it is, so a new site stays private. It stays offline until publish_site. A taken subdomain is validation_failed reason taken and may include up to three free suggestions. The site object carries the rename warning. Plans: All plans Parameters: title | string | | The site's title, shown in the browser tab and in /hive. subdomain | string | | The address: .pagehive.space. A taken name returns up to three free suggestions. theme | object or null | | Theme variables your pages print as {{ theme.name }}, such as colors and fonts. Each is text, a number or true/false. settings | object | | Site-wide settings, as an object. idempotency_key | string | | Any unique string (1-200 printable ASCII). A retry with the same key and arguments returns the first call's result instead of running again. Same as the Idempotency-Key header. Example arguments: { "title": "Team recipes", "subdomain": "team-recipes" } ### update_site https://help.pagehive.space/tool?t=update_site Update a site's title, theme, settings, visibility (private or public), status (live or offline), subdomain, agent_guide, time_zone or page_access. This does not create a version. time_zone is an IANA name used for $today in visitor actions (default UTC). agent_guide is markdown stored on the site and snapshotted when the site is published as a site template; agent_guide_edits changes part of it (send one of the two). A bot or site key cannot make a site public when the owner locked it private, or set the lock; a site key updates only its own site. An agent working as itself needs level Full and changes only title, theme, settings, time_zone and agent_guide (other fields: owner_only). Offline hides pages, /data and share links; the site still counts toward the limits, and a new site cannot go live before its first publish. A subdomain rename is limited to 3 per site lifetime, site keys included; the old address redirects for 30 days and then stops, and the site object carries this warning. A taken name returns suggestions. page_access sets path rules that keep some pages to view links, members, editors or the owner. Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. title | string | | A new title for the site. theme | object or null | | Theme variables your pages print as {{ theme.name }}. null clears them. settings | object | | Site-wide settings, as an object. status | one of: live, offline | | live serves the site; offline hides its pages, data and view links. A new site can't be set live before its first publish. offline_reason | string or null | | A note on why the site is offline, shown to visitors. null clears it. subdomain | string | | A new address. Renames are limited and the old address only redirects for 30 days, so ask the owner first. visibility | one of: private, public | | private keeps the site to the owner, members and view links; public opens it to everyone. Ask the owner before making a site public. agent_guide | string or null | | Markdown instructions for agents, stored on the site and copied with it when it becomes a site template. null clears it. agent_guide_edits | array of objects | | Find-and-replace edits on the stored agent_guide, with the same rules as patch_page edits; a failed edit changes nothing. The response then leaves the guide text out (agent_guide_chars gives its length) and reports each edit with snippets. Not with agent_guide. time_zone | string | | An IANA time zone, such as Europe/Zurich. Visitor actions use it for $today. Default UTC. page_access | array of objects | | page_access: path rules that make pages, scripts (/scripts/) and uploaded files (/files/) stricter than the site, never more open. Each rule is { path, access }. path is exact ("/notes", "/scripts/admin.js") or ends in one * that matches anything after it ("/admin/*" also covers /admin; "/scripts/admin-*"; "/*" covers everything). access: site (follows the site), link (a view or review link, a member or the owner), members (any member or the owner), editors (members with role edit or higher, or the owner), owner (the owner only). The most specific rule wins: exact beats a * rule, a longer prefix beats a shorter one. A visitor who is not allowed gets the sign-in page on link, members and editors pages, and a 404 on owner pages and on every script and file. Templates can test a link with {% if "/club" | can_open %}. Rules apply at once, without publish_site. Send the whole list each time; [] removes every rule. login_path | string or null | | A page of the site, such as /login, that visitors without access to this private site are sent to instead of the standard private-site page. Its forms post to /_pagehive/sign-in, /_pagehive/sign-in/code and /_pagehive/pin; see the guide topic login-page. null: the standard pages. Needs can_custom_login_page. outside_hosts | null | | Only null: lifts the host limit a site copied from a site template starts with, so its pages may load from and send to any host. Example arguments: { "site_id": "01a1…", "title": "Team recipes", "time_zone": "Europe/Zurich" } ### delete_site https://help.pagehive.space/tool?t=delete_site Only the owner can delete a site. This tool always fails with validation_failed and reason owner_only. The owner deletes it in /hive by typing the subdomain. Deleting there removes the site and its files and reserves the subdomain for 30 days. Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. Example arguments: { "site_id": "01a1…" } ### read_page https://help.pagehive.space/tool?t=read_page Read one page from the draft. By default it returns source only: the template, page CSS, global CSS, the partials the template includes and the scripts it references, without rendered HTML or element ids. files limits that set (template, page_css, global_css, partials, scripts, partial:name, script:name); rendered true adds the rendered HTML; elements bot returns the ids the bot wrote, and elements all adds the automatic ph- ids. comments is open (default) or all; a site key gets none. Every response carries version, the current draft version, which patch_page accepts as base_version. One response holds at most 32000 characters of file text and is never cut off without a cursor: next holds the arguments for the following read (file, offset, version, plus rendered, files and elements when set) and is null at the end; elements_offset then continues until elements_next is null. A follow-up without the first read's rendered or files can fail with validation_failed reason unknown_file, and one after the draft changed fails with reason page_changed. Page and template reference: https://pagehive.io/docs/bot-guide.md (section Pages and partials). Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. path | string | required | The page to read, such as /. comments | one of: open, all | | open (default) returns open comment threads on the page; all also returns resolved ones. files | string or array of string | | Limits the files returned: template, page_css, global_css, partials, scripts, or one by name such as partial:header. rendered | boolean | | true adds the page as rendered HTML. elements | one of: none, bot, all | | bot returns the element ids your agent wrote; all also returns the automatic ones. file | string | | The file to continue reading, from next in an earlier response. offset | integer | | Where to continue in that file, from next. limit | integer | | The most characters of file text to return. elements_offset | integer | | Where to continue the element list, from elements_next. version | integer | | The draft version from next. If the draft changed in between, the call fails with page_changed. cursor | string | | next_cursor from the previous page of this tool, to read the next page. Send the other arguments as on the first call; leave out the paging arguments the cursor replaces. Example arguments: { "site_id": "01a1…", "path": "/" } ### read_pages https://help.pagehive.space/tool?t=read_pages Read the source of up to 10 draft pages in one call: for each page its template, page CSS, global CSS, the partials it includes and the scripts it references, as read_page returns them without rendered HTML, element ids or comments. files limits that set for every page, as in read_page. All pages share one budget of 32000 characters of file text, filled in the order of paths; a page that does not fit has complete false and next, the arguments (path included) for read_page to continue it. version is the current draft version, which patch_page accepts as base_version; an optional version must equal it, or the call fails with validation_failed reason page_changed. A path that is not a page on the draft fails the whole call with not_found, details.index and details.path. Repeated paths are read once. Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. paths | array of string | required | Up to 10 pages to read, such as / and /about, in the order to read them. files | string or array of string | | Limits the files returned for every page, as in read_page. version | integer | | The draft version you expect. If the draft changed, the call fails with page_changed. Example arguments: { "site_id": "01a1…", "paths": [ "/", "/about" ] } ### read_redirects https://help.pagehive.space/tool?t=read_redirects Read the site's redirect list: source as stored and rules with line, from, to and status. version is draft (default) or live. source is null when the site has no list. Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. version | one of: draft, live | | draft (default) or live: which redirect list to read. Example arguments: { "site_id": "01a1…" } ### put_page https://help.pagehive.space/tool?t=put_page Create or replace one page in the draft. Visitors do not see it until publish_site. Omit source to leave the template. Omit page_css to leave that page's stylesheet. null or "" removes it. Creating a page requires source. Send at least one of source, page_css, title, or settings. If the files, title, and settings match the current draft, nothing is written and the response is the current version with noop true. version.live is true when that version is the one visitors currently see. The HTML is the page; a script only enhances it. Liquid output inside an inline script, an on* attribute, script src, base href, iframe srcdoc or an unquoted attribute is refused; scripts belong in /scripts files, with values passed in data-* attributes. patch_page changes part of a page. Pages load live public data with and bp.query. Page and template reference: https://pagehive.io/docs/bot-guide.md (section Pages and partials). Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. path | string | required | The page's address on the site, such as / or /about. title | string | | The page title, used when the template doesn't set its own. source | string | | The full page template: HTML with Liquid, optionally starting with a data block between --- lines. settings | object | | description, and og: { title, description, image, image_alt, site_name, url, type } for link previews. A /path image or url becomes a full address. With no image the Pagehive card shows; image "none" shows none. page_css | string or null | | CSS for this page only. null removes it. message | string | | A short note for the version history, like a commit message. verbose | boolean | | true lists every file and page of the draft. Otherwise the response lists only what this call changed, plus draft totals. Example arguments: { "site_id": "01a1…", "path": "/about", "title": "About", "source": "{% include 'header' %}

About us

" } ### patch_page https://help.pagehive.space/tool?t=patch_page Find-and-replace in the draft. Edits apply in order, each to one file: template, page_css, css, global_css, script or partial. If any edit fails, nothing is written. By default each find must match exactly once after the earlier edits: no match is find_not_found, more than one is find_ambiguous, and a count other than expect_count is count_mismatch. These errors give the edit index and, when only line endings, whitespace or one space differ, a hint that names match loose. Visitors see the edit after a publish; publish true publishes that draft in the same call, and the response then adds visibility, published_at, last_published_at and, on a private site's first publish, its one-time link (view_link, or review_link when review is on). Each applied edit reports replaced and snippets of a few lines around each replacement (omitted counts those past the cap). files and pages list only what changed (change: created, replaced, or deleted), with draft totals; verbose true lists the whole draft. Page and template reference: https://pagehive.io/docs/bot-guide.md (section Pages and partials). Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. path | string | | The page. Top level, next to edits. Required when an edit targets template or page_css; a path on an edit is dropped and the call fails with validation_failed reason path_required. message | string | | A short note for the version history, like a commit message. verbose | boolean | | true lists every file and page of the draft. Otherwise the response lists only what this call changed, plus draft totals. base_version | integer | | When set, must equal the current draft version, or the call fails with validation_failed reason version_conflict and the current version. publish | boolean | | true publishes the draft once the edits apply, even with no text change (a new version; last_published_at updates). A failed edit does not publish. edits | array of objects | required | The find-and-replace edits, applied in order. Each names file, find and replace; a script or partial edit also names it. idempotency_key | string | | Any unique string (1-200 printable ASCII). A retry with the same key and arguments returns the first call's result instead of running again. Same as the Idempotency-Key header. Example arguments: { "site_id": "01a1…", "path": "/", "edits": [ { "file": "template", "find": "

Welcome

", "replace": "

Welcome back

" } ] } ### write_files https://help.pagehive.space/tool?t=write_files Write several draft files in one call: up to 20 pages, partials, stylesheets, scripts, robots.txt, sitemap.xml and the redirect list together, stored as one draft version. Each item in files has kind (page, partial, css, script, robots, sitemap, or redirects) and source; a page also has path and optional title, settings and page_css; a partial or script has name; css has path for a page stylesheet or name "global" for the global stylesheet. delete lists items to remove by kind and path or name, including uploaded files as kind asset with name (or path asset:name); a deleted page also loses its page stylesheet. files and delete together hold at most 20 items. Files merge onto the draft; the same checks as put_page and the other writes apply, and if any item fails nothing is written. Optional base_version must equal the current draft version, or the call fails with validation_failed reason version_conflict. The whole call counts as one API call. Visitors see nothing until publish_site, unless publish true publishes this version in the same call. files and pages in the response list only what changed (change: created, replaced, or deleted), with draft totals; verbose true lists the whole draft. Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. files | array of objects | | Up to 20 pages, partials, stylesheets, scripts, robots.txt, sitemap.xml or the redirect list. Each has kind and source, plus path or name. delete | array of objects | | Files to remove, each by kind and path or name. kind asset with name removes an uploaded file. base_version | integer | | The draft version number you read last. If the draft has changed since, the call fails with version_conflict instead of overwriting. publish | boolean | | true publishes the draft in the same call, once the change is applied. message | string | | A short note for the version history, like a commit message. verbose | boolean | | true lists every file and page of the draft. Otherwise the response lists only what this call changed, plus draft totals. Example arguments: { "site_id": "01a1…", "files": [ { "kind": "partial", "name": "header", "source": "
Home
" }, { "kind": "css", "name": "global", "source": "body{font-family:system-ui}" } ] } ### put_partial https://help.pagehive.space/tool?t=put_partial Create or replace one partial. Pages include it with {% include 'name' %}. Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. name | string | required | The partial's name. Pages include it with {% include 'name' %}. source | string | required | The partial's HTML with Liquid. message | string | | A short note for the version history, like a commit message. verbose | boolean | | true lists every file and page of the draft. Otherwise the response lists only what this call changed, plus draft totals. Example arguments: { "site_id": "01a1…", "name": "footer", "source": "
© Team recipes
" } ### put_stylesheet https://help.pagehive.space/tool?t=put_stylesheet Create or replace the one global stylesheet. Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. source | string | required | The site's global CSS, used by every page. It replaces the current stylesheet. message | string | | A short note for the version history, like a commit message. verbose | boolean | | true lists every file and page of the draft. Otherwise the response lists only what this call changed, plus draft totals. Example arguments: { "site_id": "01a1…", "source": ":root{--ink:#1b1b1b}body{color:var(--ink)}" } ### put_page_css https://help.pagehive.space/tool?t=put_page_css Create or replace one page stylesheet. It loads after the global stylesheet. Page and template reference: https://pagehive.io/docs/bot-guide.md (section Pages and partials). Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. path | string | required | The page the CSS belongs to. source | string | required | CSS for that page only. message | string | | A short note for the version history, like a commit message. verbose | boolean | | true lists every file and page of the draft. Otherwise the response lists only what this call changed, plus draft totals. Example arguments: { "site_id": "01a1…", "path": "/about", "source": "h1{letter-spacing:-.02em}" } ### put_script https://help.pagehive.space/tool?t=put_script Create or replace one script file in the draft. It is not parsed as Liquid and is not visible until publish_site. The page keeps the script tag the bot wrote and loads /scripts/name. Live public rows come from /_pagehive/bp.js and bp.query. Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. name | string | required | The script's file name, such as menu.js. Pages load it from /scripts/. source | string | required | The JavaScript. message | string | | A short note for the version history, like a commit message. verbose | boolean | | true lists every file and page of the draft. Otherwise the response lists only what this call changed, plus draft totals. Example arguments: { "site_id": "01a1…", "name": "menu.js", "source": "document.querySelector('#menu')?.addEventListener('click',()=>document.body.classList.toggle('open'))" } ### upload_file https://help.pagehive.space/tool?t=upload_file Upload one file to the draft: an image (png, jpg, jpeg, gif, webp, avif, svg, ico), a PDF, a font (woff2, woff, ttf, otf) or a short video (mp4, webm). Send one of source_url or content_base64. With source_url, Pagehive downloads a file at a public https link (up to max_file_bytes, 3 redirects, 60 seconds). content_base64 carries the bytes in the call, a third larger than the file: a 93 KB PNG is tens of thousands of output tokens. create_upload_url uploads a file from a local disk without the bytes passing through a call. Pages link the file as /files/name, for example ..., after publish_site. The bytes must match the extension. The same name again replaces the file. Counts against max_file_bytes and the account's storage. Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. name | string | required | The file name. It is served at /files/ after the next publish. source_url | string | | An https link to fetch the file from. Use this for a file that is already online. content_base64 | string | | The file's bytes as base64. Only for small files: use create_upload_url for anything bigger. message | string | | A short note for the version history, like a commit message. verbose | boolean | | true lists every file and page of the draft. Otherwise the response lists only what this call changed, plus draft totals. Example arguments: { "site_id": "01a1…", "name": "logo.png", "source_url": "https://example.com/logo.png" } ### upload_files https://help.pagehive.space/tool?t=upload_files Upload up to 20 files to the draft as one draft version. Each item in files has name and one of source_url (a public https link Pagehive downloads) or content_base64; the same names, types and checks as upload_file apply. Downloads run two at a time; all files together may hold at most 100 MB. If any item fails nothing is written, and details.errors lists each failure with its index, field, reason and message. A name already on the site is replaced. Optional base_version must equal the current draft version, or the call fails with validation_failed reason version_conflict. The whole call counts as one API call. uploaded in the response lists each file (name, url, content_type, size_bytes, sha256) in the order sent; files lists what changed in the draft, as for every write. Visitors see nothing until publish_site. Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. files | array of objects | required | Up to 20 files, each { name, source_url } or { name, content_base64 }. Names follow upload_file. base_version | integer | | Optional: the current draft version. Another number fails with version_conflict. message | string | | A short note for the version history, like a commit message. verbose | boolean | | true lists every file and page of the draft. Otherwise the response lists only what this call changed, plus draft totals. Example arguments: { "site_id": "01a1…", "files": [ { "name": "hero.webp", "source_url": "https://example.com/hero.webp" }, { "name": "team.jpg", "source_url": "https://example.com/team.jpg" } ] } ### create_upload_url https://help.pagehive.space/tool?t=create_upload_url Upload a file from your own disk without its bytes passing through a tool call. Returns a single-use upload_url for curl -sS -X PUT --data-binary @ "". The cost is a few hundred tokens whatever the file size, up to max_file_bytes. No API key and no Content-Type are needed. The link works once and expires after 10 minutes (expires_at). Optional sha256 (64 hex, as sha256sum prints it) makes it accept only that exact file. The curl reply is the same as upload_file's, and the file shows as /files/name after publish_site. Same names and types as upload_file. Needs a shell that can reach pagehive.io; upload_file with source_url fetches a file that is already online. Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. name | string | required | The file name. It is served at /files/ after the next publish. sha256 | string | | Optional checksum (64 hex characters, as sha256sum prints it). The link then accepts only that exact file. message | string | | A short note for the version history, like a commit message. idempotency_key | string | | Any unique string (1-200 printable ASCII). A retry with the same key and arguments returns the first call's result instead of running again. Same as the Idempotency-Key header. Example arguments: { "site_id": "01a1…", "name": "intro.mp4" } ### create_upload_urls https://help.pagehive.space/tool?t=create_upload_urls Upload up to 20 files from your own disk as one draft version, without their bytes passing through a tool call. files lists each { name, sha256? } (same names and types as upload_file; sha256 as sha256sum prints it pins the exact file). Returns batch_id, commit_by and one single-use upload_url per file for curl -sS -X PUT --data-binary @ ""; no API key and no Content-Type are needed. The links expire after 10 minutes (links_expire_at). A sent file is staged, not yet in the draft: the curl reply says staged true and lists the batch's files still missing. Then commit_uploads with batch_id, before commit_by (an hour after this call), writes every file as one draft version. All files of a batch together may hold at most 100 MB; each counts against max_file_bytes, and staged files count against the account's storage until the batch is committed or closes. Needs a shell that can reach pagehive.io. Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. files | array of objects | required | Up to 20 files, each { name } or { name, sha256 }. Names follow upload_file; sha256 (64 hex characters, as sha256sum prints it) makes that link accept only that exact file. idempotency_key | string | | Any unique string (1-200 printable ASCII). A retry with the same key and arguments returns the first call's result instead of running again. Same as the Idempotency-Key header. Example arguments: { "site_id": "01a1…", "files": [ { "name": "hero.webp" }, { "name": "intro.mp4" } ] } ### commit_uploads https://help.pagehive.space/tool?t=commit_uploads Write every file staged through one create_upload_urls batch as one draft version. Refused, writing nothing, when a file of the batch was not sent (validation_failed reason batch_incomplete, details.missing lists the names), when the batch was already committed (batch_committed) or is past commit_by (batch_expired). Optional base_version must equal the current draft version, or the call fails with validation_failed reason version_conflict. A name already on the site is replaced. uploaded in the response lists each file (name, url, content_type, size_bytes, sha256); files lists what changed in the draft, as for every write. Visitors see nothing until publish_site. Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. batch_id | string | required | The batch_id create_upload_urls returned. Every file of the batch must have been sent. base_version | integer | | Optional: the current draft version. Another number fails with version_conflict. message | string | | A short note for the version history, like a commit message. verbose | boolean | | true lists every file and page of the draft. Otherwise the response lists only what this call changed, plus draft totals. idempotency_key | string | | Any unique string (1-200 printable ASCII). A retry with the same key and arguments returns the first call's result instead of running again. Same as the Idempotency-Key header. Example arguments: { "site_id": "01a1…", "batch_id": "01a1…" } ### put_robots https://help.pagehive.space/tool?t=put_robots Publish robots.txt on this site. A private, offline, unpublished, or unreviewed site still serves Disallow: / and ignores this file. Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. source | string | required | The text of robots.txt. message | string | | A short note for the version history, like a commit message. verbose | boolean | | true lists every file and page of the draft. Otherwise the response lists only what this call changed, plus draft totals. Example arguments: { "site_id": "01a1…", "source": "User-agent: *\nAllow: /\nSitemap: https://team-recipes.pagehive.space/sitemap.xml" } ### put_sitemap https://help.pagehive.space/tool?t=put_sitemap Publish sitemap.xml. Private, offline, and unpublished sites do not serve it. Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. source | string | required | The sitemap XML. It is served at /sitemap.xml while the site is public and online. message | string | | A short note for the version history, like a commit message. verbose | boolean | | true lists every file and page of the draft. Otherwise the response lists only what this call changed, plus draft totals. Example arguments: { "site_id": "01a1…", "source": "https://team-recipes.pagehive.space/" } ### put_redirects https://help.pagehive.space/tool?t=put_redirects Replace the site's redirect list in the draft, for old addresses of a site moved to Pagehive. Visitors get it with the next publish_site; rollback_site restores the previous list. One rule per line: from to [status], separated by spaces; # starts a comment line. from starts with / and is matched without case and without a trailing slash; a part may be fixed text, fixed text around one :name (one part, such as /:page.html), or a last part * (the rest of the path, reused as :splat). to is a path on the site or a full https:// address with a fixed host, and may use the names from from. status is 301 (default), 302, 307 or 308. The query string is passed on unless to has its own. Rules apply only where nothing is published, so a page or file always wins; first matching rule wins. from cannot be under /data or /_pagehive, and rules may not loop. A bad line is 422 validation_failed with details.line; more rules than max_redirects_per_site is 403 limit_reached. warnings lists site paths that point at nothing in the draft. Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. source | string | required | The redirect rules, one per line: from to [status]. Lines starting with # are comments. message | string | | A short note for the version history, like a commit message. verbose | boolean | | true lists every file and page of the draft. Otherwise the response lists only what this call changed, plus draft totals. Example arguments: { "site_id": "01a1…", "source": "/blog/* /news/:splat\n/old-about /about 301" } ### delete_page https://help.pagehive.space/tool?t=delete_page Delete one page and its page stylesheet. Page and template reference: https://pagehive.io/docs/bot-guide.md (section Pages and partials). Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. path | string | required | The page to remove from the draft, such as /old-news. message | string | | A short note for the version history, like a commit message. verbose | boolean | | true lists every file and page of the draft. Otherwise the response lists only what this call changed, plus draft totals. Example arguments: { "site_id": "01a1…", "path": "/old-news" } ### delete_partial https://help.pagehive.space/tool?t=delete_partial Delete one partial. Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. name | string | required | The partial to remove. message | string | | A short note for the version history, like a commit message. verbose | boolean | | true lists every file and page of the draft. Otherwise the response lists only what this call changed, plus draft totals. Example arguments: { "site_id": "01a1…", "name": "footer" } ### delete_stylesheet https://help.pagehive.space/tool?t=delete_stylesheet Delete the global stylesheet. Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. message | string | | A short note for the version history, like a commit message. verbose | boolean | | true lists every file and page of the draft. Otherwise the response lists only what this call changed, plus draft totals. Example arguments: { "site_id": "01a1…" } ### delete_page_css https://help.pagehive.space/tool?t=delete_page_css Delete one page stylesheet. Page and template reference: https://pagehive.io/docs/bot-guide.md (section Pages and partials). Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. path | string | required | The page whose CSS to remove. message | string | | A short note for the version history, like a commit message. verbose | boolean | | true lists every file and page of the draft. Otherwise the response lists only what this call changed, plus draft totals. Example arguments: { "site_id": "01a1…", "path": "/about" } ### delete_script https://help.pagehive.space/tool?t=delete_script Delete one script file. Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. name | string | required | The script to remove. message | string | | A short note for the version history, like a commit message. verbose | boolean | | true lists every file and page of the draft. Otherwise the response lists only what this call changed, plus draft totals. Example arguments: { "site_id": "01a1…", "name": "menu.js" } ### delete_file https://help.pagehive.space/tool?t=delete_file Delete one uploaded file from the draft. Pages that still link /files/name get a 404 for it after publish. Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. name | string | required | The uploaded file to remove, such as logo.png. message | string | | A short note for the version history, like a commit message. verbose | boolean | | true lists every file and page of the draft. Otherwise the response lists only what this call changed, plus draft totals. Example arguments: { "site_id": "01a1…", "name": "logo.png" } ### delete_robots https://help.pagehive.space/tool?t=delete_robots Remove a published robots.txt. A public reviewed site then serves the default Allow file. Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. message | string | | A short note for the version history, like a commit message. verbose | boolean | | true lists every file and page of the draft. Otherwise the response lists only what this call changed, plus draft totals. Example arguments: { "site_id": "01a1…" } ### delete_sitemap https://help.pagehive.space/tool?t=delete_sitemap Remove a published sitemap.xml. Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. message | string | | A short note for the version history, like a commit message. verbose | boolean | | true lists every file and page of the draft. Otherwise the response lists only what this call changed, plus draft totals. Example arguments: { "site_id": "01a1…" } ### delete_redirects https://help.pagehive.space/tool?t=delete_redirects Remove the site's redirect list from the draft. Old addresses show not found again after the next publish_site. Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. message | string | | A short note for the version history, like a commit message. verbose | boolean | | true lists every file and page of the draft. Otherwise the response lists only what this call changed, plus draft totals. Example arguments: { "site_id": "01a1…" } ### list_site_versions https://help.pagehive.space/tool?t=list_site_versions List site versions, newest first. limit defaults to 200 and cannot exceed 200. before, a version number, reads the next older page. next_before is set when another page remains. Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. limit | integer | | How many versions to return. Default and maximum 200. before | integer | | A version number. Returns older versions than that one, to page back. cursor | string | | next_cursor from the previous page of this tool, to read the next page. Send the other arguments as on the first call; leave out the paging arguments the cursor replaces. Example arguments: { "site_id": "01a1…", "limit": 10 } ### get_site_version https://help.pagehive.space/tool?t=get_site_version Get one version and its file list. Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. number | integer | required | The version number to read, from list_site_versions. Example arguments: { "site_id": "01a1…", "number": 12 } ### diff_site_versions https://help.pagehive.space/tool?t=diff_site_versions Diff two versions. added, removed and changed list files; changed text files include a unified diff. pages lists pages whose title or settings (og image, description, ...) changed, with before and after, since those changes leave the files as they were. Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. from | integer | required | The older version number. to | integer | required | The newer version number. Example arguments: { "site_id": "01a1…", "from": 11, "to": 12 } ### publish_site https://help.pagehive.space/tool?t=publish_site Publish the draft atomically. Visitors see every file in that version together. A new site stays offline until this call, which sets it live and queues the first review. visibility may be private or public; omit it and the site stays as it is, so a new site stays private. The response includes visibility. published_at is the first publish time and does not change later. last_published_at updates on every publish. A publish with no file changes still stores a new version and updates last_published_at. A locked private site cannot be made public: that call is 403 validation_failed reason visibility_locked, details.published is false, and files on that call are not written. A private site's first publish returns a one-time link in view_link (review_link when review is on) and next_steps. Optional files are included in the same version. kind may be page, partial, css, script, robots, sitemap, or redirects. Optional delete removes items by kind and path or name, as in write_files, including uploaded files (kind asset). Files merge onto the draft unless replace is true. files and delete have no count limit (write_files takes 20), so a whole site goes live as one version in one call. Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. message | string | | A short note for the version history, like a commit message. verbose | boolean | | Default true: list every file and page of the version. false lists only what this call changed, plus draft totals. replace | boolean | | true makes files the whole draft, so anything not listed is removed. By default files merge onto the draft. visibility | one of: private, public | | Set private or public while publishing. Leave it out to keep the site as it is. Ask the owner before making a site public. files | array of objects | | Files to include in the same version, in the same shape as write_files. No count limit (write_files takes 20), so a whole site goes live in one call. delete | array of objects | | Files to remove in the same version, each by kind and path or name. idempotency_key | string | | Any unique string (1-200 printable ASCII). A retry with the same key and arguments returns the first call's result instead of running again. Same as the Idempotency-Key header. Example arguments: { "site_id": "01a1…", "message": "New about page" } ### rollback_site https://help.pagehive.space/tool?t=rollback_site Copy an older version into a new version. History stays linear. Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. version | integer | required | The version number to go back to. It becomes the new live version. message | string | | A short note for the version history, like a commit message. verbose | boolean | | Default true: list every file and page of the version. false lists only what this call changed, plus draft totals. idempotency_key | string | | Any unique string (1-200 printable ASCII). A retry with the same key and arguments returns the first call's result instead of running again. Same as the Idempotency-Key header. Example arguments: { "site_id": "01a1…", "version": 11, "message": "Back to the old menu" } ### list_templates https://help.pagehive.space/tool?t=list_templates List site templates with an approved version, each at its latest approved version with the outside_hosts its pages reach. Pagehive staff review every version before it is listed. The publisher also sees its own templates that wait for review, with review_status. A site template is a frozen copy of a whole site, not a Liquid page template. Plans: All plans Parameters: none Example arguments: {} ### get_template https://help.pagehive.space/tool?t=get_template Read one site template: its approved versions with notes and preview path, the latest agent guide, and each version's review (status, reviewed_at, and the report: scripts, outside_hosts, linked_hosts, dynamic_urls, values, owner_actions, commands, warnings). The publisher also sees pending and rejected versions and the reviewer's note. Plans: All plans Parameters: slug | string | required | The template's slug, from list_templates. Example arguments: { "slug": "dev-board" } ### get_template_changes https://help.pagehive.space/tool?t=get_template_changes Return the version notes from from_version through to_version, including both. The copied site does not update itself. The notes describe changes to apply by hand. Plans: All plans Parameters: slug | string | required | The template's slug, from list_templates. from_version | integer | required | The older template version. to_version | integer | required | The newer template version. Example arguments: { "slug": "dev-board", "from_version": 2, "to_version": 3 } ### create_site_from_template https://help.pagehive.space/tool?t=create_site_from_template Copy a published site template into this account in one call. Uses the snapshot, never the live source. Requires an account-wide full key, the same as create_site. A site key is refused. The response includes agent_guide, which also stays on the new site, and the copied actions with name, table, kind, and access. Copied actions start off. update_action turns them on. The copy may reach only the template's outside_hosts (returned in the response) until the owner lifts that limit on /hive. Omit version to copy the latest approved version; a version that isn't approved is refused unless you published the template. Plans: All plans Parameters: slug | string | required | The template's slug, from list_templates. version | integer | | The template version to copy. Leave it out for the latest approved one. title | string | | The new site's title. subdomain | string | | The new site's address: .pagehive.space. table_prefix | string | | Only when this account already has a copy: the prefix for the copy's tables and values, e.g. devboard2_. Left out, a copy whose names are taken gets the first free one, and the response lists what was renamed. idempotency_key | string | | Any unique string (1-200 printable ASCII). A retry with the same key and arguments returns the first call's result instead of running again. Same as the Idempotency-Key header. Example arguments: { "slug": "dev-board", "title": "Our board", "subdomain": "our-board" } ### clear_sample_data https://help.pagehive.space/tool?t=clear_sample_data Delete every sample row on this site. Rows that are not marked sample stay. A site key can clear its own site. Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. Example arguments: { "site_id": "01a1…" } ## Tools: Data ### list_values https://help.pagehive.space/tool?t=list_values List values on this account, including read_access and site_id: the site that owns the value, or null for an account value that every site reads. A secret value is listed with value: null and value_hidden: true; read it with get_value. read_access: public = pages, page scripts (bp.value) and GET /data/_values can read it. private (default) = pages can print it ({{ values.key }}) when the site owns it or has a share, but scripts and the public JSON can't. secret = for tokens, API keys and passwords: stored encrypted, never on a page, in a script, in the public JSON or in a site template; a page that uses it is refused at save; only get_value returns the content, and list_values shows the key only. Plans: All plans Parameters: site_id | string | | Only for a site shared with you (whoami lists them): read or change that site's data as its member. Example arguments: { "site_id": "01a1…" } ### get_value https://help.pagehive.space/tool?t=get_value Get one value by key, including read_access. This is the only call that returns a secret value's content. read_access: public = pages, page scripts (bp.value) and GET /data/_values can read it. private (default) = pages can print it ({{ values.key }}) when the site owns it or has a share, but scripts and the public JSON can't. secret = for tokens, API keys and passwords: stored encrypted, never on a page, in a script, in the public JSON or in a site template; a page that uses it is refused at save; only get_value returns the content, and list_values shows the key only. Plans: All plans Parameters: key | string | required | The value's key. Example arguments: { "key": "opening_hours" } ### get_values https://help.pagehive.space/tool?t=get_values Get up to 100 values by key in one call, in the order asked, including read_access. Like get_value, it returns a secret value's content. missing lists the keys that don't exist or that this key can't read. read_access: public = pages, page scripts (bp.value) and GET /data/_values can read it. private (default) = pages can print it ({{ values.key }}) when the site owns it or has a share, but scripts and the public JSON can't. secret = for tokens, API keys and passwords: stored encrypted, never on a page, in a script, in the public JSON or in a site template; a page that uses it is refused at save; only get_value returns the content, and list_values shows the key only. Plans: All plans Parameters: keys | array of string | required | Up to 100 value keys. Keys that don't exist come back in missing. Example arguments: { "keys": [ "opening_hours", "today_total" ] } ### upsert_value https://help.pagehive.space/tool?t=upsert_value Create or replace a value by key. site_id on create puts it on that site; without it an account key puts it on the account, which every site reads, and a site key on its own site. move_value moves it later. read_access: public = pages, page scripts (bp.value) and GET /data/_values can read it. private (default) = pages can print it ({{ values.key }}) when the site owns it or has a share, but scripts and the public JSON can't. secret = for tokens, API keys and passwords: stored encrypted, never on a page, in a script, in the public JSON or in a site template; a page that uses it is refused at save; only get_value returns the content, and list_values shows the key only. template_safe defaults to false; only a template-safe value is copied into a site template, and a secret value can't be template-safe. The response hides a secret's content. Plans: All plans Parameters: key | string | required | The value's key. Pages read it as {{ values.key }}. type | string | | The value's type: text, number, date, money or percentage. value | string or number | required | A string for text and date values, a JSON number for number, money and percentage values. format | object | | How a number shows: currency, decimals, unit and compact. read_access | one of: public, private, secret | | public: pages, scripts and the public JSON can read it. private (default): only pages can print it. secret: for tokens and keys, never shown anywhere. template_safe | boolean | | true lets the value be copied into a site template. Default false; a secret can't be template-safe. site_id | string | | The id of the site. list_sites gives the ids of your sites. Example arguments: { "key": "opening_hours", "type": "text", "value": "Mon–Fri 9:00–17:00", "read_access": "public", "site_id": "01a1…" } ### upsert_values https://help.pagehive.space/tool?t=upsert_values Create or replace up to 100 values in one call and one transaction. Each item has the fields of upsert_value. If any item fails, nothing is written and details.errors lists each failure by index. site_id puts new values on that site unless an item sets its own. The response lists each saved value in the order sent, plus created and updated counts, and hides a secret's content. Plans: All plans Parameters: values | array of objects | required | Up to 100 items, each with the fields of upsert_value: key, type (required for a new key), value, and optional format, read_access, template_safe and site_id. site_id | string | | Puts new values on this site, unless an item sets its own site_id. Example arguments: { "site_id": "01a1…", "values": [ { "key": "opening_hours", "type": "text", "value": "Mon–Fri 9:00–17:00" }, { "key": "today_total", "type": "number", "value": 42 } ] } ### move_value https://help.pagehive.space/tool?t=move_value Move a value between the account and a site, or from one site to another. Full account keys only. to_site_id is a site id from list_sites, or null for the account. Every site reads account values in a page render, but only the owner and sites with a share read a site value, and a site key never reads account values. The move is refused, and nothing changes, when pages on other sites read it and keep_access is not true. keep_access true gives the old owner a read_write share and those other sites a read share. A share to the new owner is removed. dry_run true returns the plan without moving. Plans: All plans Parameters: key | string | required | The value to move to another site. to_site_id | string or null | required | A site id from list_sites, or null for the account. keep_access | boolean | | Give the old owner a read_write share and sites whose pages read it a read share. Default false. dry_run | boolean | | Return the plan without moving. Default false. Example arguments: { "key": "opening_hours", "to_site_id": "01a1…", "dry_run": true } ### delete_value https://help.pagehive.space/tool?t=delete_value Delete a value. Plans: All plans Parameters: key | string | required | The value to delete. Example arguments: { "key": "opening_hours" } ### list_tables https://help.pagehive.space/tool?t=list_tables List tables on this account, or with site_id, the tables of a site shared with you. Each table has site_id: the site that owns it, or null for an account table that every site reads. Plans: All plans Parameters: site_id | string | | Only for a site shared with you (whoami lists them): read or change that site's data as its member. Example arguments: { "site_id": "01a1…" } ### get_table https://help.pagehive.space/tool?t=get_table Get one table, including its columns. Plans: All plans Parameters: name | string | required | The table's name. site_id | string | | Only for a site shared with you (whoami lists them): read or change that site's data as its member. Example arguments: { "name": "recipes" } ### list_rows https://help.pagehive.space/tool?t=list_rows Query rows in a table. query takes where, any (AND-ed with where when both are set), search, sort, page, limit, group_by, or include, fields to return only some columns, member (a member id from list_members) for one site member's rows, and visibility (shared, private or all) on an own_and_shared table. Each row has _member: the member it belongs to, or null. _created_by and _updated_by say who wrote it: { type (bot, owner, member, visitor, reviewer or system), agent_id for an agent, name }, with the current name. Sort keeps the first occurrence of each column and allows at most 5 distinct columns. Plans: All plans Parameters: name | string | required | The table to read. query | object | | Query options: where, search, sort, limit, page and the others query_table takes. fields | array of string | | Column keys to return, such as ["slug","title"]. id, position, revision, created_at, updated_at, _member, _created_by and _updated_by always come back. Filters, search and sort can use other columns. Not with group_by. It keeps long text columns out of the response. site_id | string | | Only for a site shared with you (whoami lists them): read or change that site's data as its member. Example arguments: { "name": "recipes", "fields": [ "title", "minutes" ] } ### get_row https://help.pagehive.space/tool?t=get_row Get one row by id, with revision (goes up by one each time its values change) and created_by / updated_by as stored. _created_by and _updated_by say who wrote it: { type (bot, owner, member, visitor, reviewer or system), agent_id for an agent, name }, with the current name. A row a visitor wrote has type visitor with the actor and action. Plans: All plans Parameters: name | string | required | The table. id | string | required | The row's id. site_id | string | | Only for a site shared with you (whoami lists them): read or change that site's data as its member. Example arguments: { "name": "recipes", "id": "01a1…" } ### query_table https://help.pagehive.space/tool?t=query_table Run a data query: where, any (AND-ed with where when both are set), search, sort, paging, group_by, include, fields, and visibility. The SQL runs in Postgres. Sort keeps the first occurrence of each column and allows at most 5 distinct columns. On an own_and_shared table a member or site agent reads the shared rows plus its own private rows; visibility narrows that to shared or private. Plans: All plans Parameters: table | string | required | The table to query. where | object | | Filters by column, all of which must match: a value, or one operator such as {"gte": 10}, {"contains": "x"} or {"empty": true}. any | array of objects | | A list of where filters, of which at least one must match. search | string | | Text to find in the table's searchable columns. sort | string or array of string | | A column to sort by; a leading - sorts descending. A list sorts by several columns. limit | number | | How many rows to return per page. page | number | | The page number, starting at 1. cursor | string | | The next_cursor from the previous page, to continue. group_by | string | | A column to group by. Returns one row per value with its count and the aggregates you ask for. Not with fields. sum | array of string | | Columns to add up, per group or for the whole result. avg | array of string | | Columns to average. min | array of string | | Columns to return the smallest value of. max | array of string | | Columns to return the largest value of. include | array of string | | Relation columns whose linked rows to include in each row. fields | array of string | | Column keys to return, such as ["slug","title"]. id, position, revision, created_at, updated_at, _member, _created_by and _updated_by always come back. Filters, search and sort can use other columns. Not with group_by. It keeps long text columns out of the response. visibility | one of: shared, private, all | | Only for an own_and_shared table. shared: rows that belong to no member. private: your own private rows (the owner: every member's). all, the default: everything you may read, so the shared rows plus your own. site_id | string | | Only for a site shared with you (whoami lists them): read or change that site's data as its member. Example arguments: { "table": "recipes", "where": { "vegetarian": true, "minutes": { "lte": 30 } }, "sort": "minutes", "limit": 20, "fields": [ "title", "minutes" ] } ### list_table_changes https://help.pagehive.space/tool?t=list_table_changes List recent row changes for a table, newest first. Plans: All plans Parameters: name | string | required | The table. limit | number | | How many changes to return. before | string | | Returns changes older than this point, to page back. Use the value the previous page gave. cursor | string | | next_cursor from the previous page of this tool, to read the next page. Send the other arguments as on the first call; leave out the paging arguments the cursor replaces. Example arguments: { "name": "recipes", "limit": 20 } ### create_table https://help.pagehive.space/tool?t=create_table Create a typed table. Columns are stored as JSON and do not change the Postgres schema. site_id makes it a site table of that site; without it an account key creates an account table, which every site of the account reads, and a site key a table of its own site. move_table moves it later. read_access public is what bp.query and GET /data/ can read. A private table returns table_private. members_own (site tables only) gives each signed-in member of a private site only their own rows in the browser, and a page render never reads it. site (site tables only) follows the site: whoever may open it reads every row and file in the browser (on a private site: the owner, members, view links and the PIN; on a public site: anyone), for group sites. own_and_shared (site tables only) keeps shared rows next to each member's and site agent's private rows: they read the shared rows plus their own, the owner reads all, and a page render never reads it. owner_browser true (private site tables only) lets the owner's own visit read it there too, for owner dashboards; view links, members and strangers still get table_private. Visitor actions need a site table. Plans: All plans Parameters: name | string | required | The table's name, in lowercase letters, digits and underscores. site_id | string | | The id of the site. list_sites gives the ids of your sites. read_access | one of: public, private, members_own, site, own_and_shared | | Who can read it in the browser: public, private (pages only), members_own, site, or own_and_shared. See the table guides. owner_browser | boolean | | On a private site table: true lets the owner's own visit read it in the browser, for owner dashboards. columns | array of objects | | The columns, each with key and type (text, number, date, datetime, select, checkbox, relation or file), plus options such as required or sortable. show_last_updated | boolean | | true lets pages show when the table last changed. show_authors | boolean | | true: pages (GET /data, bp.query, Liquid) also get _created_by / _updated_by on each row. Agents show by agent name; on a private site people show by name (never an email), on a public site as Owner or Member. Default false. idempotency_key | string | | Any unique string (1-200 printable ASCII). A retry with the same key and arguments returns the first call's result instead of running again. Same as the Idempotency-Key header. Example arguments: { "name": "recipes", "site_id": "01a1…", "read_access": "public", "columns": [ { "key": "title", "type": "text", "required": true }, { "key": "minutes", "type": "number" }, { "key": "vegetarian", "type": "checkbox" } ] } ### update_table https://help.pagehive.space/tool?t=update_table Update a table's columns, read access, last-updated stamp, or name. Column changes bump schema_version. Set read_access public before a page calls bp.query, members_own so each member of the site reads only their own rows, site so everyone the site lets in reads every row and file, or own_and_shared for shared rows plus each member's and site agent's private rows. Only the owner moves a table out of members_own or own_and_shared. owner_browser true keeps a private site table private but lets the owner's own visit read it with bp.query. A rename repoints relation columns that name the table; a page that still names it blocks the rename. Plans: All plans Parameters: name | string | required | The table to change. new_name | string | | A new name for the table. read_access | one of: public, private, members_own, site, own_and_shared | | Who can read it in the browser: public, private (pages only), members_own, site, or own_and_shared. owner_browser | boolean | | On a private site table: true lets the owner's own visit read it in the browser. columns | array of objects | | The full new list of columns. Send every column you want to keep. show_last_updated | boolean | | true lets pages show when the table last changed. show_authors | boolean | | true: pages (GET /data, bp.query, Liquid) also get _created_by / _updated_by on each row. Agents show by agent name; on a private site people show by name (never an email), on a public site as Owner or Member. Default false. site_id | string | | Only for a site shared with you (whoami lists them): read or change that site's data as its member. Example arguments: { "name": "recipes", "columns": [ { "key": "title", "type": "text", "required": true, "searchable": true }, { "key": "minutes", "type": "number", "sortable": true }, { "key": "vegetarian", "type": "checkbox" } ] } ### move_table https://help.pagehive.space/tool?t=move_table Move a table between the account and a site, or from one site to another, keeping its rows, ids and history. Full account keys only. to_site_id is a site id from list_sites, or null for the account. Every site reads account data in a page render, but only the owner and sites with a share read site data, and a site key never reads account data. The move is refused, and nothing changes, when the table has read_access members_own or site, or owner_browser true, and goes to the account, when rows belong to members of the old site, when visitor actions of the old site use it, or when pages on other sites read it and keep_access is not true. keep_access true gives the old owner a read_write share and those other sites a read share. A share to the new owner is removed. dry_run true returns the plan (blockers, shares, warnings) without moving. Plans: All plans Parameters: name | string | required | The table to move to another site. to_site_id | string or null | required | A site id from list_sites, or null for the account. keep_access | boolean | | Give the old owner a read_write share and sites whose pages read it a read share. Default false. dry_run | boolean | | Return the plan without moving. Default false. Example arguments: { "name": "recipes", "to_site_id": "01a1…", "dry_run": true } ### create_row https://help.pagehive.space/tool?t=create_row Insert one row. A client-supplied id is ignored. This does not create a page version. Set sample true for a row a site template should copy. clear_sample_data removes those rows later. The row carries _created_by and _updated_by (who wrote it), so the row needs no column for the writer's name. On an own_and_shared table a new row is shared; a member or site agent passes private true to keep it to itself (and the owner). Plans: All plans Parameters: name | string | required | The table. row | object | required | The new row's column values, by column key. private | boolean | | Only for a member or site agent writing to an own_and_shared table: true keeps the new rows to you (and the owner). Default false: shared with every member and agent of the site. Rows that already exist keep their owner. site_id | string | | Only for a site shared with you (whoami lists them): read or change that site's data as its member. idempotency_key | string | | Any unique string (1-200 printable ASCII). A retry with the same key and arguments returns the first call's result instead of running again. Same as the Idempotency-Key header. Example arguments: { "name": "recipes", "row": { "title": "Lentil soup", "minutes": 35, "vegetarian": true } } ### update_row https://help.pagehive.space/tool?t=update_row Update part of a row. Null clears a column. This does not create a page version. The row's revision goes up when its values change, so a visitor edit drawn from the old revision is refused with a conflict. Plans: All plans Parameters: name | string | required | The table. id | string | required | The row's id. row | object | required | The columns to change. Columns you leave out keep their value; null clears one. site_id | string | | Only for a site shared with you (whoami lists them): read or change that site's data as its member. Example arguments: { "name": "recipes", "id": "01a1…", "row": { "minutes": 30 } } ### set_row_member https://help.pagehive.space/tool?t=set_row_member Give a row to a member of the site that owns the table (member_id from list_members), or take it back with member_id null. A row a member added through a member action already belongs to them. members_own tables and member_own actions go by this. Values and revision stay as they are. Plans: All plans Parameters: name | string | required | The table. id | string | required | The row's id. member_id | string or null | required | The member who should own the row, from list_members. null makes it a shared row. Example arguments: { "name": "tasks", "id": "01a1…", "member_id": "01a1…" } ### bulk_upsert_rows https://help.pagehive.space/tool?t=bulk_upsert_rows Upsert many rows in one transaction, by id or by a key column. A row that matches an existing id or key is merged like update_row: only the columns sent change, null clears a column, and required columns are checked only for new rows. The default response is a summary: count, inserted, updated, ids of newly inserted rows in request order, and errors. Updated rows are not in ids. return_rows true returns the full row objects. Validation errors write nothing. The error message counts errors and rows separately, for example 2 errors in 1 row failed validation, and details include error_count and row_count. Plans: All plans Parameters: name | string | required | The table. key | string | | A unique column to match rows on, such as slug. Without it, rows match on id. rows | array of objects | required | The rows. A row that matches is updated with the columns sent; the others are added. return_rows | boolean | | true returns the full rows instead of a summary. private | boolean | | Only for a member or site agent writing to an own_and_shared table: true keeps the new rows to you (and the owner). Default false: shared with every member and agent of the site. Rows that already exist keep their owner. site_id | string | | Only for a site shared with you (whoami lists them): read or change that site's data as its member. idempotency_key | string | | Any unique string (1-200 printable ASCII). A retry with the same key and arguments returns the first call's result instead of running again. Same as the Idempotency-Key header. Example arguments: { "name": "recipes", "key": "title", "rows": [ { "title": "Lentil soup", "minutes": 35 }, { "title": "Shakshuka", "minutes": 25 } ] } ### delete_table https://help.pagehive.space/tool?t=delete_table Delete a table and its rows. The change log is kept. Plans: All plans Parameters: name | string | required | The table to delete, with all its rows. Example arguments: { "name": "recipes" } ### delete_row https://help.pagehive.space/tool?t=delete_row Delete one row. A member's agent or site agent deletes rows of an own_and_shared table: the shared ones and its own, never another member's. Plans: All plans Parameters: name | string | required | The table. id | string | required | The id of the row to delete. site_id | string | | Only for a site shared with you (whoami lists them): read or change that site's data as its member. Example arguments: { "name": "recipes", "id": "01a1…" } ### bulk_delete_rows https://help.pagehive.space/tool?t=bulk_delete_rows Delete many rows by id in one transaction. A missing id writes nothing. The response has deleted_count and ids, the rows removed. The error message counts errors and rows separately, and details include error_count and row_count. A member's agent or site agent deletes rows of an own_and_shared table: the shared ones and its own, never another member's. Plans: All plans Parameters: name | string | required | The table. ids | array of string | required | The ids of the rows to delete. site_id | string | | Only for a site shared with you (whoami lists them): read or change that site's data as its member. Example arguments: { "name": "recipes", "ids": [ "01a1…", "01a1…" ] } ### list_data_shares https://help.pagehive.space/tool?t=list_data_shares List shares of tables and values. An account key sees every share. A site key sees shares of data it owns and shares granted to its site. Plans: All plans Parameters: none Example arguments: {} ### create_data_share https://help.pagehive.space/tool?t=create_data_share Share one table or one value from its owning site to another site. Only an agent that works as the owner on the whole hive, with no sites left out, may call this; site keys and agents limited to some sites are refused (reason share_whole_hive). access is read or read_write. The same pair again updates access. Plans: All plans Parameters: to_site_id | string | required | The site that gets access. table | string | | The table to share. Give either table or value. value | string | | The value to share, by key. Give either table or value. access | one of: read, read_write | required | read lets the other site read it; read_write also lets it change it. idempotency_key | string | | Any unique string (1-200 printable ASCII). A retry with the same key and arguments returns the first call's result instead of running again. Same as the Idempotency-Key header. Example arguments: { "to_site_id": "01a1…", "table": "recipes", "access": "read" } ### revoke_data_share https://help.pagehive.space/tool?t=revoke_data_share Revoke a share. Only an agent that works as the owner on the whole hive, with no sites left out, may call this; site keys and agents limited to some sites are refused (reason share_whole_hive). Plans: All plans Parameters: id | string | required | The share's id, from list_data_shares. Example arguments: { "id": "01a1…" } ## Tools: Visitors and members ### list_members https://help.pagehive.space/tool?t=list_members List a site's members: id, email, status (invited or active), role (view, edit, build or publish), agents (whether their own agent may work on the site), and when they were invited and accepted. The id is the member id pages get from /_pagehive/me and rows carry as _member. Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. Example arguments: { "site_id": "01a1…" } ### list_site_pins https://help.pagehive.space/tool?t=list_site_pins List a site's PINs: id (the member id rows carry as _member), label, role, created_at and last_used_at. Never the PIN itself. A PIN lets people open a site by typing it on the site's PIN page, with no account or email. A site has up to 10 PINs, each 6 to 12 digits. Each PIN is a member without an email: role view sees the site, role edit also runs member actions and uploads, and rows it adds belong to that PIN. label says what it is for (Grandparents). The PIN works until it is removed. Typing it keeps that device signed in for 30 days, or until the browser closes when the visitor ticks the public computer box. The PIN is stored hashed and is never shown again after it is set. Removing a PIN ends the visits that came from it. Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. Example arguments: { "site_id": "01a1…" } ### invite_member https://help.pagehive.space/tool?t=invite_member Invite an email address to a site, public or private. On a public site everyone else stays a visitor; the member signs in to get their role. role is view (default: sees the site), edit (also runs member actions and commands), build (their own agent may also change the draft and the site's tables) or publish (and publish). agents true lets their own agent work on the site within that role; it defaults to false. Pagehive emails a link that works once for 7 days; the link is never returned to you. Inviting the same address again sends a new link and voids the old one, and changes the role when you send one. The account must be claimed and have can_invite_members. Limits: max_members_per_site and max_member_invites_per_day (in get_account). Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. email | string | required | The address to invite. Pagehive emails a link that works once, for 7 days. role | one of: view, edit, build, publish | | view (default) sees the site; edit also runs member actions and commands; build lets their agent change the draft and tables; publish also lets it publish. agents | boolean | | true lets the member's own agent work on the site within that role. Default false. idempotency_key | string | | Any unique string (1-200 printable ASCII). A retry with the same key and arguments returns the first call's result instead of running again. Same as the Idempotency-Key header. Example arguments: { "site_id": "01a1…", "email": "sam@example.com", "role": "edit" } ### update_member https://help.pagehive.space/tool?t=update_member Change a member's role (view, edit, build, publish) and/or agents (whether their own agent may work on the site, within the role). Send at least one. It applies from the member's next click or call. Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. member_id | string | required | The id of the member, from list_members. role | one of: view, edit, build, publish | | The new role: view, edit, build or publish. agents | boolean | | true lets the member's own agent work on the site within their role; false stops it. Example arguments: { "site_id": "01a1…", "member_id": "01a1…", "role": "view" } ### create_site_pin https://help.pagehive.space/tool?t=create_site_pin Create a PIN for a site. pin: 6 to 12 digits as a string, not a run like 123456 and not the same as another PIN of the site. role view (default) or edit, label (what it is for). A PIN lets people open a site by typing it on the site's PIN page, with no account or email. A site has up to 10 PINs, each 6 to 12 digits. Each PIN is a member without an email: role view sees the site, role edit also runs member actions and uploads, and rows it adds belong to that PIN. label says what it is for (Grandparents). The PIN works until it is removed. Typing it keeps that device signed in for 30 days, or until the browser closes when the visitor ticks the public computer box. The PIN is stored hashed and is never shown again after it is set. Removing a PIN ends the visits that came from it. Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. pin | string | required | The PIN people type to open the site. role | one of: view, edit | | view (default) sees the site; edit also runs member actions. label | string or null | | A name to tell PINs apart, such as Kitchen tablet. idempotency_key | string | | Any unique string (1-200 printable ASCII). A retry with the same key and arguments returns the first call's result instead of running again. Same as the Idempotency-Key header. Example arguments: { "site_id": "01a1…", "pin": "482913", "role": "view", "label": "Kitchen tablet" } ### update_site_pin https://help.pagehive.space/tool?t=update_site_pin Change a PIN's role, label (null clears it), or the PIN itself (pin). A new PIN signs out everyone who used the old one. Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. pin_id | string | required | The id of the PIN, from list_site_pins. pin | string | | A new PIN. role | one of: view, edit | | The new role: view or edit. label | string or null | | A new label. null clears it. Example arguments: { "site_id": "01a1…", "pin_id": "01a1…", "label": "Hallway tablet" } ### revoke_member https://help.pagehive.space/tool?t=revoke_member Remove a member from a site. Their invite link and any open visit stop working at once. Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. member_id | string | required | The id of the member, from list_members. Example arguments: { "site_id": "01a1…", "member_id": "01a1…" } ### remove_site_pin https://help.pagehive.space/tool?t=remove_site_pin Remove a PIN. Everyone who opened the site with it loses access at once; other PINs keep working. Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. pin_id | string | required | The id of the PIN, from list_site_pins. Example arguments: { "site_id": "01a1…", "pin_id": "01a1…" } ### read_site_data https://help.pagehive.space/tool?t=read_site_data On a site shared with you (whoami lists them): read what you, as its member, can read on the site. With table: that table's rows, the same as bp.query on the site (public and site tables, and only your own rows of a members_own table); query takes where, search, sort, limit, mine, fields and the other query options. Without table: the site's public values. On the account's own sites, list_rows and list_values read the same data. Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. table | string | | The table to read. Leave it out to read the site's public values. query | object | | Query options: where, search, sort, limit, fields, mine and the others query_table takes. fields | array of string | | With table: column keys to return. id, position, revision, created_at, updated_at, _member, _created_by and _updated_by always come back. It keeps long text columns out of the response. Example arguments: { "site_id": "01a1…", "table": "tasks" } ### run_action https://help.pagehive.space/tool?t=run_action On a site shared with you with role edit or higher: run one of the site's visitor actions as its member, exactly as a click on the page would. Only actions with access member_own, member_all or public. A row you add belongs to you; member_own changes only your own rows. Send row for an existing row, values for edit and add, value and params for set and fixed, revision to refuse a stale edit, and request_id (a UUID) so a retry runs once. list_actions lists them. Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. name | string | required | The action to run, from list_actions. row | string | | The id of the row to change. Not for an add action. values | object | | Column values for an edit or add action. value | string | | The chosen value for a set action. params | object | | Parameters for a set or fixed action, such as a number of days. revision | integer | | The row's revision you read. If the row changed since, the action is refused. request_id | string | | A UUID of your own, so a retry of the same call runs only once. idempotency_key | string | | Any unique string (1-200 printable ASCII). A retry with the same key and arguments returns the first call's result instead of running again. Same as the Idempotency-Key header. Example arguments: { "site_id": "01a1…", "name": "toggle_done", "row": "01a1…" } ### list_agents https://help.pagehive.space/tool?t=list_agents List the agents that work on a site as themselves (its own member there): member_id (the _member of the rows they keep private), agent_id (the same agent on every site), name, role (from the agent's level), allowed_tools, and their active site keys for this site (prefix and last use, never the secret). Agents that work as the owner are not members and are not listed. The owner adds and changes agents in /hive under Agents. Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. Example arguments: { "site_id": "01a1…" } ### list_actions https://help.pagehive.space/tool?t=list_actions List visitor actions on a site, with kind, access and spec (column, allowed, when, set, params, fields). access off means the action refuses every caller, including the owner. Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. Example arguments: { "site_id": "01a1…" } ### action_log https://help.pagehive.space/tool?t=action_log Read the visitor action log, newest first. Each entry has op (insert for an add, delete for a delete, update otherwise), the actor, member_id, agent ({ agent_id, name }, the current name) when one of your agents ran it as its member, the share link id when one was used, a hash of the IP, the row, and the old and new value of each column whose value changed. A delete keeps the row's old values. before and before_id page backward. Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. limit | integer | | How many entries to return. before | string | | Returns entries older than this time, to page back. before_id | string | | With before: the last entry's id, so entries at the same time aren't skipped. cursor | string | | next_cursor from the previous page of this tool, to read the next page. Send the other arguments as on the first call; leave out the paging arguments the cursor replaces. Example arguments: { "site_id": "01a1…", "limit": 50 } ### action_events https://help.pagehive.space/tool?t=action_events Visitor actions since an exclusive timestamp, oldest first. since, and since_id from next when a page is full, page through the events, for mirroring calls into the site's own log. Each event has op (insert, update, or delete), the row and its changes; row values can hold text that visitors typed. Omit since to start at the oldest stored entry. The worker deletes rows older than action_log_retention_days (90), so this feed only reaches back that far. When the owner set the site's Webhook with visitor actions on, each action is also sent to it. Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. since | string | | Returns events after this time, oldest first. Leave it out to start at the oldest stored event. since_id | string | | With since: the id from next, so events at the same time aren't skipped. limit | integer | | How many events to return. cursor | string | | next_cursor from the previous page of this tool, to read the next page. Send the other arguments as on the first call; leave out the paging arguments the cursor replaces. Example arguments: { "site_id": "01a1…", "limit": 50 } ### define_action https://help.pagehive.space/tool?t=define_action Define a visitor action on one table this site owns. kind toggle flips a checkbox. kind set writes a text or select column to one of at least two allowed values, and when can set other columns from the chosen value ($today, $now, null, or a literal). kind fixed writes constant values, or a bounded integer param that sets a date to $today plus that many days. kind edit lets the visitor write the columns in fields (each a column key, or { column, max_length, min, max, required }); set adds fixed writes. kind add creates a row from fields plus set; every required column must be in one of them. kind delete removes the row. edit, add and delete are only off, owner, share, member_own, or member_all, never public. A relation column can be an edit or add field (the visitor sends a row id) or a fixed row id in set or when, only with access off, owner, member_own or member_all; the related table must be one this site reads, and a call checks that the caller may read the linked row. access defaults to off. public needs can_public_actions. member_own lets members of the site with role edit add rows, which then belong to them, and change or delete only their own rows; member_all lets them change any row. An owner visit passes both. An account key may name an account-owned table no other site uses; it then moves to this site, so the action can come before the page. A site key cannot take account tables. Other sites' tables and shared tables are refused. Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. name | string | required | The action's name. Pages run it by this name. table | string | required | The table the action changes. kind | one of: toggle, set, fixed, edit, add, delete | required | toggle flips a checkbox; set picks one allowed value; fixed writes set values; edit, add and delete change, create or remove a row. access | one of: off, owner, share, share_review, public, member_own, member_all | | Who may run it: off (default), owner, share, share_review, public, member_own or member_all. edit, add and delete are never public. column | string or null | | The column a toggle or set action writes. allowed | array of string | | For set: the values a visitor may choose, at least two. when | object | | For set: other columns to write for each chosen value, such as a done date. set | object | | Fixed column values written every time the action runs. params | array of objects | | For fixed: a bounded number a visitor chooses, such as days from today, each with name, min, max and column. fields | array of string or object | | For edit and add: the columns a visitor may write, each a column key or an object with limits. idempotency_key | string | | Any unique string (1-200 printable ASCII). A retry with the same key and arguments returns the first call's result instead of running again. Same as the Idempotency-Key header. Example arguments: { "site_id": "01a1…", "name": "toggle_done", "table": "tasks", "kind": "toggle", "column": "done", "access": "public" } ### update_action https://help.pagehive.space/tool?t=update_action Update a visitor action. Send the fields that change. Changing kind starts the spec over from the fields you send. An action that is already public can be edited when can_public_actions is off, so access can move away from public. Setting access to public needs the right. Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. name | string | required | The action to change. table | string | | A different table for the action. kind | one of: toggle, set, fixed, edit, add, delete | | A new kind: toggle, set, fixed, edit, add or delete. access | one of: off, owner, share, share_review, public, member_own, member_all | | Who may run it: off, owner, share, share_review, public, member_own or member_all. column | string or null | | The column a toggle or set action writes. allowed | array of string | | For set: the values a visitor may choose. when | object | | For set: other columns to write for each chosen value. set | object | | Fixed column values written every time the action runs. params | array of objects | | For fixed: bounded numbers a visitor chooses, each with name, min, max and column. fields | array of string or object | | For edit and add: the columns a visitor may write. Example arguments: { "site_id": "01a1…", "name": "toggle_done", "access": "member_all" } ### delete_action https://help.pagehive.space/tool?t=delete_action Delete a visitor action. Logged calls stay in the action log. Rows the action wrote or added stay. Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. name | string | required | The action to delete. Example arguments: { "site_id": "01a1…", "name": "toggle_done" } ## Tools: Sharing ### list_site_domains https://help.pagehive.space/tool?t=list_site_domains List a site's custom domains: hostname, role (primary serves the site, redirect sends to it), status (pending, active, dns_lost, disabled, off with off_by owner or plan), the last check's error and the DNS records to create. Needs can_custom_domains. Plans: Basic and up (basic, pro, ultra) Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. Example arguments: { "site_id": "01a1…" } ### add_site_domain https://help.pagehive.space/tool?t=add_site_domain Connect a custom domain (such as www.example.com or example.com) to a site. It stays pending until verify_site_domain passes. domain.dns.records holds the DNS records the owner adds at their DNS host: a TXT record proving they own the domain, and a CNAME (or A and AAAA for an apex). role: primary serves the site and is the default for its first domain; redirect sends visitors to the primary address. Apex and www are two separate domains. Needs can_custom_domains. Plans: Basic and up (basic, pro, ultra) Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. hostname | string | required | The domain to connect, such as www.example.com. The apex and www are separate domains. role | one of: primary, redirect | | primary serves the site; redirect sends visitors to the primary address. The first domain is primary by default. idempotency_key | string | | Any unique string (1-200 printable ASCII). A retry with the same key and arguments returns the first call's result instead of running again. Same as the Idempotency-Key header. Example arguments: { "site_id": "01a1…", "hostname": "www.example.com" } ### verify_site_domain https://help.pagehive.space/tool?t=verify_site_domain Check a domain's DNS after the owner created the records: the TXT record, and that every A and AAAA answer is Pagehive. Passing makes it active and the site is served on it at once (a primary domain also takes over from the pagehive.space address, which redirects there). Failing returns error with what is missing; DNS changes can take minutes, so wait and try again rather than in a loop. Plans: Basic and up (basic, pro, ultra) Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. domain_id | string | required | The id of the domain, from list_site_domains. Example arguments: { "site_id": "01a1…", "domain_id": "01a1…" } ### update_site_domain https://help.pagehive.space/tool?t=update_site_domain Change a domain: role primary or redirect (making one primary turns the site's current primary into a redirect), and/or on false to turn it off or on true to turn it back on. Off keeps the domain and its settings but stops serving it and frees its place under the plan's custom domain limit: the way to fit a plan with fewer custom domains without removing them. Only a verified domain can be turned off. Turning on needs room on the plan. Plans: Basic and up (basic, pro, ultra) Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. domain_id | string | required | The id of the domain, from list_site_domains. role | one of: primary, redirect | | primary serves the site; redirect sends visitors to the primary address. on | boolean | | false turns a verified domain off: kept with its settings, not served, not counted towards the plan. true turns it back on when the plan has room. Example arguments: { "site_id": "01a1…", "domain_id": "01a1…", "role": "primary" } ### remove_site_domain https://help.pagehive.space/tool?t=remove_site_domain Disconnect a custom domain from a site. Removing the primary makes the next domain primary (a live one first, reported as promoted); with none left, the site is served on its pagehive.space address again. Plans: Basic and up (basic, pro, ultra) Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. domain_id | string | required | The id of the domain, from list_site_domains. Example arguments: { "site_id": "01a1…", "domain_id": "01a1…" } ### list_site_links https://help.pagehive.space/tool?t=list_site_links List share links for a site with expires_at and expired. Tokens are not included. Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. Example arguments: { "site_id": "01a1…" } ### list_site_keys https://help.pagehive.space/tool?t=list_site_keys List site keys for one site. Each item includes label, prefix, and dates. Secrets are never returned. Requires an account key. A site key cannot list keys. Only the owner makes, rotates and revokes keys, on the agent's page in /hive under Agents. Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. Example arguments: { "site_id": "01a1…" } ### create_site_link https://help.pagehive.space/tool?t=create_site_link Create a view link for a site. The URL is shown once. The link does nothing while the site is offline. Review links are paused. It expires after 30 days; expires_at (ISO date) sets another date, and null means no expiry. Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. kind | one of: view | required | view: a link that opens a private site without signing in. expires_at | string or null | | When the link stops working, as an ISO date. Default 30 days; null means it never expires. idempotency_key | string | | Any unique string (1-200 printable ASCII). A retry with the same key and arguments returns the first call's result instead of running again. Same as the Idempotency-Key header. Example arguments: { "site_id": "01a1…", "kind": "view", "expires_at": "2026-12-31T23:00:00Z" } ### revoke_site_link https://help.pagehive.space/tool?t=revoke_site_link Revoke a share link. Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. link_id | string | required | The id of the view link, from list_site_links. Example arguments: { "site_id": "01a1…", "link_id": "01a1…" } ### seo_scan https://help.pagehive.space/tool?t=seo_scan SEO and sharing report for one site: what search engines and link previews (X, LinkedIn, WhatsApp, Slack) get from each page, and what to improve. It renders each page as visitors get it and checks: search (indexing, robots.txt, sitemap.xml, title and description per page, canonical, viewport, lang), sharing (own share image vs the Pagehive card, image format, size and shape for large previews, og and twitter tags), structure (one h1, heading order, image alt text, broken same-site links, pages nothing links to, redirects that lead to a 404 or another redirect, structured data, own icon) and weight (heavy HTML, images over 1 MB). findings have check, group (search, sharing, structure, weight), severity high/medium/low/info, title, detail, where, evidence and a fix (a tool and how). A private site is checked as if it were public. draft true checks the draft instead of the live version; a site never published is always checked from its draft. Share images on other websites are not fetched. max_seo_scan_pages per scan is a tier setting; pages past it are counted, not checked. 10 scans per site per hour. The last report shows on the site's SEO tab in /hive. Not for a site shared with you. Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. draft | boolean | | true checks the draft instead of the live version. Example arguments: { "site_id": "01a1…" } ## Tools: Your Hive and security ### scan_site https://help.pagehive.space/tool?t=scan_site Security report for one site: what it exposes, and to whom. reach lists every live page, table, value, uploaded files, action and command with how a stranger, a view link, a member and a member's own agent get in (page, data, own_rows, source, rows, run). findings are grouped as site, pages, data and shared, each with severity high/medium/low/info, where, evidence and a fix. Private keeps a table out of /data, not out of pages: a page that prints a private table shows its rows. It reads the settings and structure of the live version, never row values or what pages and files contain. probe true also runs a live test: requests to the site as a stranger, through the code that serves visitors, that confirm pages, data and files open or stay closed as the settings say, that draft-only and earlier-version pages do not open, and that the response headers are in place (live_test; not counted as visits; status codes and headers only). Without probe, not_checked lists those checks. 10 scans per site per hour. Not for a site shared with you. Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. probe | boolean | | true also tests the live site as a stranger would, to confirm pages, data and files open or stay closed as the settings say. Example arguments: { "site_id": "01a1…" } ## Tools: Agent commands ### list_commands https://help.pagehive.space/tool?t=list_commands List agent commands on a site, and agent: whether the owner set the site's Webhook (connected, kind claude, chatgpt, grok_bot, or generic), which events it gets (events: actions, comments, form_submits, ticks, review_done; send_actions repeats actions), and pending_commands. Read-only: only the owner sets the Webhook and its events, in /hive (site › Webhook). Never the URL or token. Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. Example arguments: { "site_id": "01a1…" } ### command_queue https://help.pagehive.space/tool?t=command_queue Read a site's command queue. status pending (default) lists commands waiting for the agent, oldest first, including claims whose lease ran out. in_progress, done, failed, or all list newest first. Each entry has the instruction to carry out, which keeps {input} where the visitor's text goes, the visitor's text as input, who sent it, and the page. claim_command and complete_command take a command through the queue. Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. status | one of: pending, in_progress, done, failed, all | | pending (default) lists commands waiting for the agent, oldest first. The others list newest first. limit | integer | | How many commands to return. Example arguments: { "site_id": "01a1…", "status": "pending" } ### define_command https://help.pagehive.space/tool?t=define_command Define an agent command a page can send with data-bp-command or bp.command. kind fixed (default) needs instruction, up to 2000 characters; put {input} where the visitor's text goes and set input_max (default 500, up to 4000). kind free has no instruction: the owner's own text is sent, up to input_max (default 2000), and access is owner or off. access is off (default), owner, share (the owner and anyone with a share link), or member (the owner and members of the site with role edit). deliver_to owner (default) sends every call to the owner's agent; sender sends a member's call to that member's own agent instead (queued for that member; it starts the agent the member connected for this site on their own /hive, else it waits for that agent's next run), and anyone else's to the owner's. Commands are never public. At most 50 per site. Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. name | string | required | The command's name. Pages send it by this name. kind | one of: fixed, free | | fixed (default) sends your instruction with the visitor's text in it; free sends the owner's own text. label | string or null | | The text on the command's button. instruction | string or null | | For fixed: what the agent should do, up to 2000 characters. Put {input} where the visitor's text goes. input_max | integer or null | | The most characters a visitor may type. access | one of: off, owner, share, member | | Who may send it: off (default), owner, share (also share links) or member (also members with role edit). Never public. deliver_to | one of: owner, sender | | owner (default) sends calls to the owner's agent; sender sends a member's call to that member's own agent. idempotency_key | string | | Any unique string (1-200 printable ASCII). A retry with the same key and arguments returns the first call's result instead of running again. Same as the Idempotency-Key header. Example arguments: { "site_id": "01a1…", "name": "summarize", "kind": "free", "label": "Ask the agent", "instruction": "Answer this question about the recipes table: {input}", "access": "owner" } ### update_command https://help.pagehive.space/tool?t=update_command Update an agent command. Send the fields that change. name renames it. Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. name | string | required | The command to change. new_name | string | | A new name for the command. kind | one of: fixed, free | | fixed or free. label | string or null | | The text on the command's button. null clears it. instruction | string or null | | For fixed: what the agent should do. Put {input} where the visitor's text goes. input_max | integer or null | | The most characters a visitor may type. access | one of: off, owner, share, member | | Who may send it: off, owner, share or member. deliver_to | one of: owner, sender | | owner or sender: whose agent gets the calls. Example arguments: { "site_id": "01a1…", "name": "summarize", "label": "Ask about recipes" } ### claim_command https://help.pagehive.space/tool?t=claim_command Claim a queued command so another run skips it while this one works on it. The lease lasts lease_seconds (60 to 3600, default 600); after that the command is pending again. An error with reason claimed or finished means skip it. Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. command_id | string | required | The id of the queued command, from command_queue. lease_seconds | integer | | How long the claim lasts, 60 to 3600 seconds. Default 600. After that the command is pending again. idempotency_key | string | | Any unique string (1-200 printable ASCII). A retry with the same key and arguments returns the first call's result instead of running again. Same as the Idempotency-Key header. Example arguments: { "site_id": "01a1…", "command_id": "01a1…" } ### complete_command https://help.pagehive.space/tool?t=complete_command Close a queued command as done or failed, with an optional note of up to 1000 characters that the page can show the sender. Completing a connection test marks the owner's agent connection as working. Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. command_id | string | required | The id of the queued command, from command_queue. status | one of: done, failed | required | done or failed. note | string or null | | A note of up to 1000 characters that the page can show the sender. Example arguments: { "site_id": "01a1…", "command_id": "01a1…", "status": "done", "note": "Added 3 recipes." } ### complete_commands https://help.pagehive.space/tool?t=complete_commands Close up to 50 queued commands in one call, each as done or failed with an optional note, with the same rules as complete_command. A bad argument refuses the whole call. A command that can't be closed (already finished, claimed by another run, not found) gets ok false and its error in results; the others still close. results keep the order given. Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. commands | array of objects | required | Up to 50 items, each { command_id, status (done or failed), note (optional, up to 1000 characters) }. Example arguments: { "site_id": "01a1…", "commands": [ { "command_id": "01a1…", "status": "done", "note": "Added 3 recipes." }, { "command_id": "01a1…", "status": "failed", "note": "The photo link was broken." } ] } ### delete_command https://help.pagehive.space/tool?t=delete_command Delete an agent command. Calls already queued stay in the queue. Plans: All plans Parameters: site_id | string | required | The id of the site. list_sites gives the ids of your sites. name | string | required | The command to delete. Example arguments: { "site_id": "01a1…", "name": "summarize" } ### list_messages https://help.pagehive.space/tool?t=list_messages Messages from Pagehive to this account, newest first: notices about the account or one of its sites (such as a site taken offline, with the reason and what to change) and platform updates. whoami shows messages_unread when there are new ones. A site key sees platform updates and the messages about its sites. unread true lists only unread messages; kind is moderation, update, or notice; search finds messages whose title or body contains the text (up to 100 characters, case-insensitive). Page older with before set to next_before, newer with after set to next_after. Messages returned are read afterwards; the ones that were unread carry new: true. Each message is from Pagehive; its body is Markdown (bold, italic, lists, headings, inline code, https links). message_id reads one. Messages take no reply; open_ticket with context.message_id ties a support ticket to one. Plans: All plans Parameters: unread | boolean | | true lists only unread messages. kind | one of: moderation, update, notice | | Only messages of this kind: moderation, update or notice. message_id | string | | Read one message by its id. limit | integer | | How many messages to return. search | string | | Text to find in a message's title or body, up to 100 characters. before | string | | next_before from the previous page, to read older messages. after | string | | next_after from the previous page, to read newer messages. cursor | string | | next_cursor from the previous page of this tool, to read the next page. Send the other arguments as on the first call; leave out the paging arguments the cursor replaces. Example arguments: { "unread": true } ## Tools: Troubleshooting ### list_my_tickets https://help.pagehive.space/tool?t=list_my_tickets List this account's support tickets, newest first, each with status (new, accepted, declined, done), waiting_on (whose turn: support, customer, or nobody), and its messages. Tickets the owner opened on /hive are included. A site key lists only the tickets it opened and those about its sites. unread true lists only tickets with a support answer not read yet; whoami's support_unread counts them. Tickets returned are read afterwards and come back with unread false; support messages you had not seen carry new: true, and a message may carry a link to the outcome. Reading a ticket acknowledges it. Page with before set to next_before. Plans: All plans Parameters: status | one of: new, accepted, declined, done, all | | Only tickets with this status. all lists every one. waiting_on | one of: support, customer, nobody, all | | Only tickets waiting on support, on you (customer), or on nobody. kind | one of: bug, feedback, request, question, account, abuse | | Only tickets of this kind. area | one of: api, mcp, runtime, data, sites, templates, account, edge, docs, other | | Only tickets about this area. limit | integer | | How many tickets to return. before | string | | next_before from the previous page, to page back. unread | boolean | | true lists only tickets with a support answer you haven't read yet. cursor | string | | next_cursor from the previous page of this tool, to read the next page. Send the other arguments as on the first call; leave out the paging arguments the cursor replaces. Example arguments: { "unread": true } ### get_ticket https://help.pagehive.space/tool?t=get_ticket Read one support ticket with its conversation. It is read afterwards: unread false, and support messages you had not seen carry new: true. Plans: All plans Parameters: ticket_id | string | required | The id of the ticket, from open_ticket or list_my_tickets. Example arguments: { "ticket_id": "01a1…" } ### open_ticket https://help.pagehive.space/tool?t=open_ticket Open a Pagehive support ticket. kind is bug, feedback, request, question, account (about this account, tier, or keys), or abuse (a site or account breaking the rules). Give a short title (3 to 120 characters) and a detail (up to 4,000) with what you did, what you expected, and what happened. area is api, mcp, runtime, data, sites, templates, account, edge, docs, or other. severity (critical, major, minor, trivial) is for bugs. context is an optional object, such as the tool or route, the error code, and the request id. Pagehive keys in a ticket are removed before it is stored. Any key can open a ticket, including a site key. The same kind and title within 24 hours, before any answer, adds to the earlier ticket (merged true). An account sends at most 10 tickets and replies an hour and 30 a day. When support answers, whoami shows support_unread; list_my_tickets with unread true returns the answers. Plans: All plans Parameters: kind | one of: bug, feedback, request, question, account, abuse | required | What it is: bug, feedback, request, question, account or abuse. title | string | required | A short title for the ticket. detail | string | | What happened, what you expected, and the steps to see it. area | one of: api, mcp, runtime, data, sites, templates, account, edge, docs, other | | The part of Pagehive it's about. severity | one of: critical, major, minor, trivial | | How bad it is, for a bug: critical, major, minor or trivial. site_id | string | | The id of the site. list_sites gives the ids of your sites. context | object | | Extra details as an object, such as a request id or the error you got. idempotency_key | string | | Any unique string (1-200 printable ASCII). A retry with the same key and arguments returns the first call's result instead of running again. Same as the Idempotency-Key header. Example arguments: { "kind": "bug", "title": "publish_site times out on a large site", "detail": "Three publishes in a row ended in a timeout around 30 seconds.", "area": "sites", "severity": "major", "site_id": "01a1…" } ### reply_to_ticket https://help.pagehive.space/tool?t=reply_to_ticket Reply on a support ticket, up to 4,000 characters, when you have something new: an answer to support's question, what you found, or that a fix did not work. Reading an answer already acknowledges it. The reply hands the turn to support (waiting_on support) and keeps the status. On a closed ticket (done, declined) it is a note support will see; set reopen true when the problem is back or the fix did not work, which sets the status to new (reopened true). Replies count against the same limit as new tickets. A ticket holds at most 50 messages and takes replies until 90 days after its last message; after that, open a new ticket and name the old id in context. Plans: All plans Parameters: ticket_id | string | required | The id of the ticket, from open_ticket or list_my_tickets. message | string | required | Your reply, up to 4,000 characters. Send it when you have something new, not only to say thanks. reopen | boolean | | true reopens a closed ticket, when the problem is back or a fix didn't work. idempotency_key | string | | Any unique string (1-200 printable ASCII). A retry with the same key and arguments returns the first call's result instead of running again. Same as the Idempotency-Key header. Example arguments: { "ticket_id": "01a1…", "message": "It works again since this morning. Thanks!" }