Skip to content
English
  • There are no suggestions because the search field is empty.

How to set up and use Avoma's MCP server to analyze meetings

Use your conversations and Avoma insights to do more in your AI tools. Generate better documents, build stronger strategies, and create custom reports tailored to your team.

Avoma gives teams a powerful way to capture and act on meeting insights with transcripts, summaries, action items, CRM sync, and analytics available out of the box. But for teams that want to go beyond prebuilt workflows and extract insights based on their own logic, goals, and models, the MCP server offers the next level of flexibility.

Why connect to Avoma's MCP Server?

With Avoma's MCP (Model Context Protocol) server, you can securely connect your meeting data to Claude Desktop or any AI tool that supports the MCP standard and put it to work in ways that go beyond Avoma's built-in workflows.

What you can do with Avoma’s MCP server

  1. "Create an internal champion brief for the Acme deal based on all conversations with their team."

  2. "Generate a weekly voice of the customer report from support and success conversations."

  3. "Summarize the top sales lessons from this quarter's win/loss calls and turn them into a coaching playbook."

  4. "Summarize recurring product feedback from customer calls this quarter and create a prioritized list of feature requests and bugs."

Before you begin

To use Avoma’s MCP server, you must have Node.js installed on your computer.

The setup uses npx mcp-remote, which depends on npm/npx being available.

How to check if Node.js is installed

Open Terminal (Mac) or Command Prompt / PowerShell (Windows) and run:

node -v
npm -v
npx -v

If these commands return version numbers, you’re ready to continue.

If you see "not found" or an error, you need to install Node.js:

  1. Go to nodejs.org
  2. Download the LTS version (recommended for most users)
  3. Run the installer and follow the steps
  4. Once done, close and reopen your Terminal or Command Prompt

Run node -v again to confirm it installed correctly

Step-by-step: Connect your AI tool to Avoma’s MCP server

We'll use Claude Desktop as an example here, but it will work with any AI tool that supports MCP.

Step 1: Get your Avoma API key

In Avoma, go to Settings → Organization → Developer and copy your API key. You'll need the full key string to authenticate. It follows the format CLIENT_KEY:CLIENT_SECRET.

Step 2: Add Avoma's MCP server to your config

Open Claude Desktop and go to Settings → Developer → Edit Config

Avoma MCP configuration

This will open the claude_desktop_config.json file. If it opens in your browser instead of a text editor, right-click the file and select Open With → TextEdit.

Step 3: Locate the "mcpServers" section and add the following:

{
"mcpServers": {
"avoma": {
"command": "npx",
"args": [
"mcp-remote",
"https://mcp.avoma.com/mcp",
"--header",
"Authorization: Bearer YOUR_AVOMA_API_KEY"
]
}
}
}

Replace YOUR_AVOMA_API_KEY with your actual key.

Avoma MCP API key config

Replace YOUR_AVOMA_API_KEY with your actual API key. 
Save the file and restart your AI tool (Claude in this case).

Step 3: Confirm your setup is working

Run a test query like:

  • “List all meetings from last week”
  • “Show transcript for the onboarding call with <company>”



If the tool returns results using Avoma data, your setup is good to go.

Important
Keep queries focused for faster, more accurate results. Where possible, narrow by timeframe like the last few months, meeting type such as demo, title, or company name. Broad queries that pull thousands of meetings will be slower and may time out.

Step 5: Approve access to MCP tools when prompted

When you run a request that requires Avoma data, the assistant will ask for permission to use Avoma MCP tools.

When prompted, approve access to the Avoma MCP tool.

After you approve, the requested data will be retrieved and displayed automatically. 

Enable access to Avoma MCP tools when prompted

If the tool returns results using Avoma data, your setup is good to go.

For tool-specific setup details, refer to the official MCP documentation for Claude.

Using a different AI tool? 

The steps above are specific to Claude Desktop. Other tools that support MCP will have their own configuration methods — refer to your tool's documentation for instructions on adding an MCP server.

Available MCP tools from Avoma

Here are the tools exposed via Avoma’s MCP server:

Tool

Description

Example prompt

list_meetings

Find meetings by filters

"Meetings with John last week"

get_meeting

View meeting details

"What happened in meeting #12345?"

get_meeting_transcript

Full conversation log

"Show transcript of onboarding call"

get_meeting_notes

View meeting summary

"What were key takeaways from Friday’s demo?"


Troubleshooting

If you're having trouble connecting Claude to Avoma or accessing meeting data, try the following solutions for common issues:

Issue: Claude is not recognizing your API key

Verify you copied the full key string from Settings → Organization → Developer. If the issue persists, regenerate your key from the same location.

If the key is invalid, generate a new one from Settings → Organization → Developer.

Issue: Tools are not appearing in Claude

If the Avoma tools (like list_meetings or get_meeting_notes) are not showing up in Claude, try the following steps:

  • Restart Claude Desktop after updating the configuration file
  • Check Claude’s logs for any errors related to MCP setup or permissions
  • Verify that your claude_desktop_config.json file is correctly structured and that there are no formatting issues in the JSON

References

  1. Set up your API key in your Avoma account
  2. Configure Claude Desktop with the MCP server