joinly.ai is a connector middleware designed to enable AI agents to join and actively participate in video calls. Through its MCP server, joinly.ai provides essential meeting tools and resources that empower AI agents to perform tasks and interact in real time during meetings. It supports cross-platform access to Google Meet, Zoom, and Microsoft Teams directly in the browser, and offers live interaction via voice or chat. The server includes built-in conversational flow to handle interruptions and multi-speaker dynamics, and supports Bring-your-own-LLM by working with all providers (also locally with Ollama). It uses a modular architecture for Text-to-Speech and Speech-to-Text, with providers such as Whisper/Deepgram for STT and Kokoro/ElevenLabs/Deepgram for TTS, among others. The MCP server is open-source, self-hosted, and privacy-first. Developers can run it with Docker, connect external clients, and configure multiple MCP servers via config files. Useful tools include join_meeting, leave_meeting, speak_text, send_chat_message, mute_yourself, get_transcript, get_participants, get_chat_history, and get_video_snapshot, plus a live transcript resource.