joinly-ai

joinly: Connect AI Agents to Live Meetings with MCP

Joinly is MCP middleware that lets AI agents join Google Meet, Zoom, and Teams calls to speak, chat, and use meeting context in real time. It uses Whisper for transcription and supports self-hosting with a choice of LLM providers.

ActivePythonMITAI AgentsMCPMCP Server
Stars
568
Forks
94
Stars, last 30 days
+4
Last commit
2 months ago
Contributors
5

Our take

Activerecent releases

Health score

38/100

How it is scored
Activity
48
Community
22
Issues
21
Pull requests
60

+4 stars in the last 30 days

Joinly is an actively maintained, MIT-licensed way to connect MCP agents to browser-based meetings, though project work is highly concentrated in one contributor.

Good fit if

  • You are building a Python agent and want live meeting access through MCP.
  • You can self-host the browser-based service and configure an LLM and optional speech services.
  • You want to combine meeting interaction with tools from other MCP servers.

Look elsewhere if

  • You need a meeting service with authentication on its MCP server, since the README says the server has none.
  • You need camera use, screen sharing, or participant join/leave metadata, which the roadmap lists as unfinished.
  • You require a project whose maintenance is spread across multiple contributors; most contributions currently come from one person.
All health signals
Last commit2026-09-01 (2 months ago)
Commits, last 90 days6
Releases, last 12 months3 (latest v0.5.4, 2026-09-01)
Contributors5 (top contributor: 94% of commits)
Issues closed, last 90 days0
Pull requests merged, last 90 days3 (typically merged in 0 days)
Project age17 months

Checked on 2026-10-25 with the GitHub API.

Project history

Daily GitHub snapshots since Oct 7, 2026. History covers stars and forks; health signals are not tracked historically.

Selected snapshot

Stars

573

Forks

96

Change since Nov 16, 2026: +1 stars·0 forks

Overview

Joinly is Python middleware built around an MCP server that gives AI agents access to live video meetings. It uses Whisper for speech transcription and lets agents join calls on Google Meet, Zoom, Microsoft Teams, or other platforms available through a browser, then interact by voice or chat.

It is aimed at developers who want meeting participation to be part of an agent workflow, rather than a separate note-taking step. Agents can access transcripts and meeting details, respond in real time, and use tools from other MCP servers during a call. The project is self-hostable and supports different LLM providers.

Key Features

  • Join browser-based meetings on Google Meet, Zoom, and Microsoft Teams.
  • Speak text, send chat messages, and mute or unmute through MCP tools.
  • Retrieve live transcripts, chat history, participant lists, and video snapshots.
  • Subscribe to the transcript://live resource for transcript updates with timestamps and speaker information.
  • Use Whisper for speech-to-text, with local Kokoro text-to-speech by default and optional Deepgram or ElevenLabs services.
  • Connect additional MCP servers so their tools are available to the meeting agent.
  • Run as a Dockerized client or as an MCP server for a separate client.

Use Cases

  • A developer can let a meeting agent answer questions using web-search or other MCP tools during a call.
  • A team can have an agent follow a discussion and retrieve the transcript or chat history when needed.
  • An agent can share information aloud or in meeting chat while participants are collaborating.
  • A developer building an agent client can use Joinly's MCP tools and meeting resources as its real-time call interface.

What you need

Detected in the repository

  • Python >=3.12 (from pyproject.toml)
  • A test suite and automated checks on GitHub Actions

License in plain words

MITpermissive

  • Commercial use: yes
  • Modify and redistribute: yes
  • You must keep: the copyright and license notice
  • Share your changes: no

A summary, not legal advice: the LICENSE file is what applies.

Getting Started

Create a .env file with your chosen LLM provider settings, then pull and run the Docker image with a meeting URL:

docker pull ghcr.io/joinly-ai/joinly:latest
docker run --env-file .env ghcr.io/joinly-ai/joinly:latest --client <MeetingURL>

See the README for configuration, custom clients, and development instructions.

Alternatives

  • mityu: Mityu records and transcribes meetings locally to create source-linked notes; Joinly joins live calls as an agent that can speak and chat.
ProjectLanguageLicenseStarsStatus
joinlyPythonMIT568Active
mityuRustMIT1New

Considerations

  • The project is active, with recent commits and releases, but its five contributors are not evenly distributed: one contributor accounts for most of the work. No issues were closed in the measured recent period, while pull requests were merged.
  • The MCP server has no authentication and accepts client-supplied configuration. The README advises binding it to localhost and not exposing it to a network.
  • The Docker image includes a browser and speech assets and is large. CPU use is supported; GPU acceleration is optional and requires the NVIDIA Container Toolkit and CUDA.
  • Development requires Python 3.12 or later. The README identifies camera support, screen sharing, participant metadata, and improved turn-taking as unfinished roadmap items.
  • The project uses the MIT License.

Found this useful?

Share it with someone who would like joinly.

Comparisons