MCP Server

The Record Meeting MCP (Model Context Protocol) server lets AI assistants like Claude, Cursor, Codex, ChatGPT, and other MCP-compatible clients access your recordings, transcripts, and organizations directly.

What is MCP?

Model Context Protocol is an open standard that lets AI assistants connect to external tools and data sources. By connecting your AI assistant to the Record Meeting MCP server, you can ask questions about your meetings, search transcripts, manage recordings, and more - all from within your AI workflow.

Let your agent set itself up

If you already have an AI agent open, you do not need to follow this guide by hand. Copy this line and paste it into a fresh chat:

Fetch and execute the appropriate instructions to set me up for RecordMeeting from https://recordmeeting.com/agent-setup/prompt.md

Your agent reads the instructions at that URL, connects through the sign-in flow (or asks you for an API key if its tool cannot sign in), writes the configuration for whichever tool you are using, and then verifies the connection by listing one of your recordings. The same button is available in the app under Settings, and on the agent setup page.

The rest of this page covers the manual setup, which is worth reading if you want to know exactly what your agent is changing.

Prerequisites

• A Record Meeting account with at least one recording
• An MCP-compatible client (e.g. Claude Desktop, Claude Code, Cursor, Codex, ChatGPT)
• An API key (see API Keys), only if your client cannot sign in on its own

Connecting your client

The Record Meeting MCP server is available at:

https://api.recordmeeting.com/mcp

Sign in instead of pasting a key

Most clients can connect with a sign-in flow, so you do not need an API key at all. Add the server with no headers, and the client opens a Record Meeting page in your browser where you approve the connection. Claude Code, Claude Desktop, Claude.ai, Cursor and ChatGPT all support this.

For Claude Code:

claude mcp add --transport http recordmeeting https://api.recordmeeting.com/mcp

Then run /mcp inside Claude Code, pick recordmeeting, and choose Authenticate. For Cursor, add the server without a headers block and click the login prompt next to it in Settings, then MCP. For Claude Desktop, Claude.ai and ChatGPT, add https://api.recordmeeting.com/mcp as a custom connector and follow the sign-in.

You can see and disconnect every connected agent under Settings, in the Connected agents section. The sections below use an API key instead, for clients that only support static headers.

Claude Desktop

Add the following to your Claude Desktop configuration file (claude_desktop_config.json):

{
  "mcpServers": {
    "recordmeeting": {
      "url": "https://api.recordmeeting.com/mcp",
      "headers": {
        "Authorization": "Bearer <your_api_key>"
      }
    }
  }
}

Claude Code

Run the following command in your terminal:

claude mcp add --transport http recordmeeting https://api.recordmeeting.com/mcp \
  --header "Authorization: Bearer <your_api_key>"

Cursor

Add the following to your .cursor/mcp.json file:

{
  "mcpServers": {
    "recordmeeting": {
      "url": "https://api.recordmeeting.com/mcp",
      "headers": {
        "Authorization": "Bearer <your_api_key>"
      }
    }
  }
}

Codex

Add the following to ~/.codex/config.toml (or a project-scoped .codex/config.toml):

[mcp_servers.recordmeeting]
url = "https://api.recordmeeting.com/mcp"
http_headers = { Authorization = "Bearer <your_api_key>" }

ChatGPT

ChatGPT desktop shares MCP configuration with Codex. Add the same block to ~/.codex/config.toml, or paste it under Settings → MCP servers:

[mcp_servers.recordmeeting]
url = "https://api.recordmeeting.com/mcp"
http_headers = { Authorization = "Bearer <your_api_key>" }

Note

Replace <your_api_key> with an actual API key from Settings → API Keys. See API Keys for how to create one.

Available tools

Once connected, your AI assistant can use the following tools:

Recordings

Tool	Description
list_recordings	List your recordings with optional filters (status, date range, folder, organization)
get_recording	Get details for a specific recording including download link and notes
get_transcript	Retrieve transcript segments for a recording
search_recordings	Search recordings by keyword across titles and transcripts
get_recording_by_date	Find recordings from a specific date or date range
update_recording	Update a recording’s title or meeting notes
delete_recording	Permanently delete a recording
get_share_link	Create a sharing link with configurable permissions

Organizations

Tool	Description
list_organizations	List organizations you belong to
list_folders	List folders within an organization
list_members	List members of an organization with their roles

Example prompts

Once connected, try asking your AI assistant things like:

• “What meetings did I have last week?”
• “Summarize the key decisions from yesterday’s standup”
• “Search my recordings for discussions about the product launch”
• “What action items came out of the Q1 planning meeting?”
• “Share the recording from this morning’s sync with view-only access”

Troubleshooting

“Unauthorized” or “401” errors

• Make sure your API key is valid and has not been revoked
• Check that the Authorization header uses the Bearer prefix

Tools not appearing

• Verify the server URL is exactly https://api.recordmeeting.com/mcp
• Restart your AI client after updating the configuration

No recordings returned

• Confirm your account has recordings by checking the Record Meeting dashboard
• If using an organization context, make sure your role grants access to recordings