MCP Guide

Connect your AI agent to Anchor.

One URL. OAuth sign-in. Nineteen tools for apps, tables, files, and sharing. Works with every MCP-compatible agent.

Anchor MCP server Remote · HTTP · OAuth 2.0 + PKCE
https://mcp.anchor.cc/mcp

Add this URL as a remote MCP server in your agent. Sign in with the Google account tied to your Anchor workspace when prompted.

Set up your agent Browse all 19 tools

What is MCP?

The Model Context Protocol (MCP) is the open standard for AI agent tool use — think "USB-C for AI." It defines how an agent discovers and calls tools on a remote server. Any MCP-compatible agent can talk to any MCP server without custom integrations.

The Anchor MCP server exposes nineteen tools covering every action you can take in Anchor: create and edit apps, design and query tables, upload and download files, organize folders, and manage sharing. Your agent calls these tools on your behalf; Anchor handles storage, hosting, permissions, and versioning.

One connection, every agent. No vendor lock-in. No webhooks. No custom integration code.

Connect in three steps

For most agents the whole setup takes less than a minute. The flow is the same everywhere: paste one URL, authorize with Google, done.

  1. Copy the server URL

    Your agent only needs this one value:

    https://mcp.anchor.cc/mcp

  2. Add Anchor as a remote MCP server

    Open your agent's MCP (or "connector") settings and add a new remote / HTTP MCP server pointing at the URL above. See the per-agent setup instructions for the exact path in Claude, ChatGPT, Gemini, Codex, Grok, and Copilot.

  3. Authorize with Google

    Your agent opens an OAuth window. Sign in with the Google account tied to your Anchor workspace and approve. Tokens are issued automatically and refresh on use — you never see an API key.

Supported agents

Anchor works with every MCP-compatible agent. The following are verified and have dedicated setup walkthroughs:

Any other MCP-compatible client also works — the server follows the Model Context Protocol spec exactly.

Tool reference

The Anchor MCP server exposes nineteen tools, grouped by capability. The table below is the canonical summary — agents discover the same list at runtime via tools/list, and every tool reports its own input and output schema.

Tool What it does
Navigation & discovery
documentation Return top-level context about Anchor — file types, structure, sharing model. Agents call this first to ground themselves before acting.
list_orgs List the organizations the authenticated user belongs to. Returns each org's ID, name, and creation timestamp.
ls Inspect an org, folder, file, app slug, or everything shared with you. The workhorse of navigation — resolves URLs and slugs to file/folder metadata.
Apps
create_app Create a new app from a single index.tsx React file, optional CSS, optional table-backed functions, optional file connections, and a Lucide-icon favicon. Builds server-side and auto-publishes v1.
edit_app Rename an app and/or apply surgical text edits to its source. Triggers a rebuild and creates a new version; auto-publishes if the app isn't shared, otherwise requires manual approval.
read_app Read one source file at a time from an app — index.tsx, index.css, functions.json, or favicon.json. Used to inspect code before editing.
list_app_functions List the callable backend functions an app exposes, including input parameters and output shape — everything needed to call them.
call_app_function Execute one of an app's functions with positional arguments. Returns the result as an array of row objects.
Tables
create_table Create a new table with typed columns (STRING, NUMBER, INTEGER, BOOLEAN, ENUM, DATE, DATETIME, EMAIL), constraints, and optional foreign keys. Every table includes an auto-increment record_id.
describe_table Return a table's full schema, row count, connected apps, and timestamps. Does not return row data.
query_table Execute a SQL statement (SELECT / INSERT / UPDATE / DELETE) against Anchor tables. Schema-changing statements are rejected; joins across tables in the same org are allowed.
add_table_column Add a new column to an existing table. Required columns must include a default value so existing rows can be backfilled.
delete_table_column Permanently delete a column and all its data. Deferred as a pending schema change when a published app depends on the column.
Files
file_upload_link Generate a deep link to the Anchor UI where the user can manually upload files into a folder. For OTHER file types only.
file_link Return view and download links for an OTHER file. For apps and tables, use the corresponding tools instead.
read_file Read the parsed text content of a file — PDFs, spreadsheets, CSVs, markdown, and other text-parseable formats. Supports line-range reads.
Folders & organization
mkdir Create a new folder at the org root or inside an existing folder.
mv Move or rename a file or folder within the same org. Supports rename-in-place by passing only new_name.
cp Copy a file into a folder. Apps get a rebuilt bundle; tables get a fresh duplicate of schema and data. Folders cannot be copied.

Authentication & security

Anchor's MCP server uses OAuth 2.0 with PKCE — the same flow Claude, ChatGPT, and other MCP clients already speak natively. No client secrets, no static API keys, no copy-paste tokens.

Discovery
On first request, your agent fetches /.well-known/oauth-protected-resource and /.well-known/oauth-authorization-server from mcp.anchor.cc to learn where to authorize.
Dynamic client registration
Your agent registers itself via POST /oauth/register. No pre-provisioning required.
User sign-in
The authorization endpoint redirects to Google SSO. You sign in with the same Google account you use for anchor.cc.
Tokens
Access tokens are short-lived JWTs. Refresh tokens rotate on use. Revoke access anytime from your Anchor account settings.
Authorization
Every tool call runs as you. Your agent can only see and modify content you have access to — the same permission model used in the UI.

The server implements the Model Context Protocol OAuth flow as specified by modelcontextprotocol.io — any MCP client that follows the spec will authenticate correctly.

Usage & limits

Each MCP tool call counts as one operation against your organization's monthly quota. Free organizations get 100,000 operations per month; paid plans raise the ceiling. See pricing for the full breakdown.

Large reads — such as read_file on long PDFs or query_table on wide tables — are billed as a single operation regardless of row count. Rate limits apply per-token to keep runaway agents from exhausting your quota.

FAQ

Which MCP transport does Anchor support?

Streamable HTTP — the recommended transport for remote MCP servers. Any MCP client that speaks Streamable HTTP over OAuth 2.0 works out of the box. Local stdio transports are not required or supported.

Do I need an API key?

No. Anchor issues OAuth access tokens transparently when you connect your agent. There is no API key to generate, store, or rotate.

Can one agent connection access multiple orgs?

Yes. You authorize once; the agent sees every organization your Google account belongs to. Use list_orgs to enumerate them and ls to navigate into a specific one.

Can I self-host the MCP server?

Enterprise customers can run Anchor — including the MCP server — under their own domain or inside a managed VPC. Contact us for details.

Where is the full tool spec?

Every tool's input and output schema is served by the MCP server itself — connect your agent and call tools/list to retrieve it. The Tool reference above is the same list in human-readable form.

Pick your agent and connect.

Per-agent walkthroughs for Claude, ChatGPT, Gemini, Grok, Codex, and Copilot.

View supported agents