Skip to article
Bridge

Check and control what an AI client can do with the Bridge MCP

The Bridge MCP lets an AI client like Claude, ChatGPT or Gemini reach your Bridge meetings once you authorize it. It can read meeting context, artifacts and transcripts, analyze transcripts and notes, and take actions. This page covers plan requirements on both sides, how to test the connection and how to stay in control.

What MCP is and what the Bridge server offers

The Model Context Protocol (MCP) is an open standard for connecting AI apps to outside systems. Its documentation compares it to a USB-C port: one standard plug instead of a custom integration for every pair of tools.

In practice, you can ask your usual AI client about last week's customer calls, and it reads the relevant Bridge transcripts. No more copying text between tools. The Bridge MCP server supports three kinds of work:

  • Retrieve meeting context, artifacts and transcripts.
  • Analyze transcripts and notes.
  • Take actions with meeting data, such as creating apps or updating records.

Before you connect

  • Plan: the Bridge MCP isn't available on the Starter plan. Authorization is blocked at the OAuth step.
  • Client: Claude, ChatGPT and Gemini are supported, along with other MCP-compatible clients.
  • Client plan: in Claude, custom connectors work on Free, Pro, Max, Team and Enterprise. Free allows one custom connector.
  • Teams: on Claude Team and Enterprise, an owner adds the connector for the organization first. Each member then connects.
  • Data: the OAuth step gives the client access to your Bridge meetings. Decide first whether that client and its provider should see them.

The setup in brief

In Bridge, go to Account settings → Integrations → Bridge MCP and copy the MCP URL from the banner. Paste it into your AI client's connector or server field, then finish the OAuth sign-in.

Our MCP page at meetonbridge.com/mcp has step-by-step instructions for each client.

In Claude on a Pro or Max plan, Anthropic's help center says to go to Customize → Connectors. Choose Add custom connector and paste the server URL. On Team and Enterprise, an owner has to add it first.

Check that it works, then put it to use

Start with a read-only question you can check. For example, ask for the decisions from a specific recent meeting and compare the answer with its notes in Bridge. If the client can't find the meeting, confirm the OAuth step finished and the connector is on for that chat.

Ask the client to name the meetings it used, so you can check its sources. The best requests combine meeting context with work in your AI client:

  • Draft a project update from several meetings' notes.
  • Compare what two customers asked for.
  • Pull every mention of a topic across a quarter.

Stay in control of actions

The server can act as well as read, so treat a connected client like any tool with write access. Anthropic notes that a connected custom connector lets Claude access the service and take action in it.

Review proposed changes before you approve them. Disconnect clients you no longer use. On a shared plan, agree who may add connectors.

Need just one transcript instead of ongoing access? Export it from Bridge as PDF, Word or Markdown.

Sources & further reading