joinly
Open-source middleware connecting AI agents to Google Meet, Zoom, and Teams video calls via Model Context Protocol.
Connect LLMs to live audio and chat in video calls to automate tasks like updating Notion or GitHub.
Joinly is open-source middleware that connects AI agents to video conferencing tools like Zoom, Google Meet, and Microsoft Teams. It exposes meeting controls and live audio over the Model Context Protocol (MCP), letting your custom agents listen, talk, and perform actions during meetings. You should use it if you want an AI assistant in standups or planning sessions to update project management tools in real time.
What it is
The tool runs as a local Docker container or cloud service that joins browser-based video calls as a participant. It handles speech-to-text, text-to-speech, and conversational flows like handling interruptions while routing requests to your choice of LLM. Through MCP integrations, the agent can access external tools—such as searching the web, creating GitHub issues, or editing Notion documents while the meeting is happening.
Who it's for
It suits technical project managers, Agile coaches, and engineering leads who want to automate meeting administrative tasks directly from live voice discussions. It does not suit non-technical teams looking for an out-of-the-box bot without Docker experience or those unwilling to configure API keys for LLM and speech services.
Key features
Cross-Platform Video Call Access
Joins Google Meet, Zoom, and Microsoft Teams directly through automated browser sessions.
Model Context Protocol Integration
Connects external tools like GitHub, Notion, or web search APIs to your voice agent via MCP servers.
Modular Speech and LLM Engines
Supports local speech processing with Whisper and Kokoro alongside cloud options like ElevenLabs, Deepgram, OpenAI, and Claude.
Real-Time Interruption Handling
Manages multi-speaker dynamics and conversational flow so the agent responds naturally to verbal interruptions.
Up and running in 3 steps
Our honest verdict
What's great
- Runs entirely self-hosted via Docker or locally with Ollama for full data privacy
- Supports third-party MCP servers to grant agents direct access to external developer and project tools
- Offers flexibility to mix local TTS/STT engines with cloud-based LLM providers
- Handles live voice interaction and text chat within meeting rooms simultaneously
Where it falls short
- Requires Docker container management and local port configuration to run self-hosted
- Heavy resource usage when running local browser automation and speech models together (over 2.3GB Docker image)
- Server mode lacks built-in authentication, requiring strict local binding to avoid security risks
- Voice response latency depends heavily on your selected TTS, STT, and LLM API combination
Pricing & license
Joinly is 100% open-source under the MIT license, meaning you can self-host it for free. Your only costs when self-hosting come from third-party API usage (such as OpenAI, ElevenLabs, or Deepgram) and local hardware resources. A hosted cloud version is also available at cloud.joinly.ai.
Open remote AI-training roles
879 open remote AI-training roles, updated November 15, 2026; most pay $20–$175 an hour. Zekai lists them free; the application goes to the employer’s platform.
Alternatives
These reviewed AI tools do the same job instantly — from our scored directory.
Use it, or have us set it up
FAQ
Can Joinly run completely offline?
Yes, you can pair Joinly with local speech tools like Whisper and Kokoro alongside a local LLM running via Ollama.
Which video conferencing platforms does it support?
It works with Google Meet, Zoom, Microsoft Teams, and any web-based meeting platform accessible via browser.
Is there a cloud version available if I do not want to host Docker?
Yes, Joinly offers a managed cloud service at cloud.joinly.ai for deployment without local setup.
