M
MCP Directory CURATED DIRECTORY
911 MCP servers indexed
中文
← Back to directory
C

Call.md

Community
Gateway / proxy
Turn meetings into live agent loops

This gateway aggregates or proxies other MCP servers; it only becomes useful once you connect real servers behind it.

How to install ↓ Compare GitHub source repository ↗
Category
Collaboration #55 of 55
Stars
★ 1.5k Popular
Transport
stdio (local process)
Runtime
Prebuilt binary · Node.js 18+
Credentials
API key / credential required
Last commit
20FMRS · D

Call.md is a feature-rich meeting recording and AI assistant app with MCP integration for automatic tool triggering during meetings. It offers dual-channel transcription, live metrics, and AI-generated assists. As a non-official server (actually an MCP client), it requires desktop installation and a VideoDB API key.

Strongest · Documentation 5/20 Weakest · Reliability 2/20

Reliability
2/20
Security and permissions
5/20
Maintenance
4/20
Documentation
5/20
Setup experience
4/20
Why each score
Reliability 2/20
The supplied material contains no server manifest, MCP entry point, or tool definitions. The README describes an Electron desktop app that acts as an MCP client connecting to external servers, not an MCP server under review. Startup, MCP handshake, tool list, and behavioral consistency cannot be verified, and there is no test or CI evidence. Core evidence is missing, so reliability is limited to 2.
Security and permissions 5/20
No source code is available for review. Visible risks include a curl|bash install script, a required VideoDB API key, microphone and screen-recording permissions, external network dependence, and automatic triggering of MCP tools during meetings. The README partially discloses these risks, but credential storage, least privilege, confirmation of dangerous operations, and scoping of auto-triggered tools are undocumented. No explicit red-line evidence exists, so it avoids the 0-4 band, but insufficient evidence caps the score at 5.
Maintenance 4/20
The repository is not archived and has 408 stars and 6 open issues, but the metadata lists the license as unknown, and the README badge claiming MIT is not supported by a license file in the supplied material. There is no evidence of releases, commit frequency, issue response, dependency updates, or a security-response channel, so the maintenance score is 4.
Documentation 5/20
The README is detailed for the desktop app's installation, features, data storage, and troubleshooting, but it does not document this repository as an MCP server. There are no tool parameters, MCP capability descriptions, client configuration examples, rate or cost limits, or verifiable examples. Key documentation is missing, so the score is 5.
Setup experience 4/20
There is a one-line installer for macOS and Windows and npm commands for developers, but the target is an Electron app rather than an MCP server. There is no server manifest, no mainstream client connection configuration example, and no Linux support. The path to installing this as an MCP server in a client is unclear, so the setup score is 4.

Static review · not runListed 2026-08-07

Read the FMRS scoring method →

Fit and risk

What it can accessWrites / deletes local filesUses the networkConnects to a database

Best for

  • Professionals who attend frequent meetings and need real-time assistance and automation
  • Teams looking to automate meeting documentation and follow-up actions
  • Individuals and teams who want to integrate meeting data into workflows like n8n, Zapier, or CRMs

Not for

  • Simple note-taking without need for real-time AI assistance
  • Organizations requiring fully offline operation (requires internet)
  • Users seeking an open-source, self-hosted solution (license unknown)

Required permissions

  • Microphone access for recording
  • Screen recording access for screen capture
  • VideoDB API key (requires internet)

Risks and side effects

  • Requires internet connectivity for transcription and AI features
  • Data may be transmitted to VideoDB, posing privacy concerns
  • Early-stage application, may have bugs or feature limitations
  • Requires system permissions, which could be a security concern if mishandled

Setup

Before you start

Runtime：Prebuilt binary · Node.js 18+

Install via curl -fsSL https://artifacts.videodb.io/call.md/install | bash (macOS and Windows only currently). After installation, launch the app, grant microphone and screen recording permissions, and register with a VideoDB API key.

Check that it works

Launch Call.md, register with your VideoDB API key, grant Microphone and Screen Recording permissions, then click New Meeting — live dual-channel transcription appearing within 5-10 seconds and a successful Connect in Settings → MCP Servers confirm the install works.

Troubleshooting

  1. If recording doesn't start, check microphone and screen recording permissions
  2. Verify VideoDB API key is correctly entered in settings
  3. Ensure stable internet connection for real-time transcription
  4. For development issues, run `npm run rebuild` to rebuild native modules

Things to try

Once connected, you can ask your AI assistant things like:

  • Add an stdio-transport MCP server under Settings → MCP Servers and connect to it
  • Summarize the meeting I just recorded with key points and action items
  • During the meeting, detect information needs from the conversation and call my connected MCP tools automatically
  • Export the latest recording as Markdown with the full transcript and metrics

Use cases

Real-time transcription and note-taking during meetings
Getting AI suggestions and coaching nudges during calls
Automatically triggering MCP tools for on-the-fly information retrieval
Generating structured meeting summaries with action items

Overview

Call.md is a desktop application that records, transcribes, and analyzes meetings with real-time AI intelligence. It captures dual-channel audio (you vs them), provides live assists and coaching nudges, and generates post-meeting summaries with action items. As an MCP client, it automatically triggers MCP tools during meetings based on conversation context, displaying results inline. Powered by VideoDB, it requires a VideoDB API key and internet connectivity.

Similar servers

Compare with other MCP servers

Source revision ba53ebed3643 Data synced 2026-10-12 Read the FMRS scoring method