joinly provides an interface for integrating AI agents directly into virtual meetings. These agents actively participate by writing in the chat, performing real-time tasks, and offering interactive assistance during sessions. The platform enables agents to handle post-meeting actions like drafting follow-up emails or generating summaries.
Funding
Funding not disclosed
Founders
Product
Problem
Many virtual meetings require participants to manually switch between conversation, note‑taking, web searches, and document creation, leading to interruptions, duplicated effort, and delayed follow‑up actions. Existing tools often provide only passive transcription or static note‑taking without real‑time AI assistance.
Solution
joinly offers a platform that embeds AI agents directly into video conference environments, allowing the agents to interact via chat and voice, execute tasks on demand, and automate post‑meeting deliverables. Leveraging the Model Context Protocol (MCP), joinly connects the meeting host to external AI services and data sources, enabling secure two‑way communication without custom scripting. During a call, the AI can fetch web information, draft slides, update collaborative boards, and generate summaries or follow‑up emails, then speak the results back to participants. Users can run the solution locally via a Docker‑based MCP server or use a hosted cloud instance for instant access. The platform’s extensible architecture lets organizations plug in additional MCP‑compatible tools (e.g., Notion, Tavily) to tailor the agent’s capabilities to specific workflows.
Target Audience
Primary users are knowledge workers, product managers, and remote teams who run frequent virtual meetings and need AI‑augmented assistance for real‑time information retrieval, content creation, and post‑meeting automation.
Features
- MCP‑driven integration layer that provides real‑time chat messaging, text‑to‑speech, mute/unmute control, and live transcript access within the meeting UI
- Built‑in toolset for web search, presentation drafting, Miroboard updates, and automated follow‑up email or ticket creation
- Extensible plugin model allowing connection to any external MCP server (e.g., knowledge bases, search engines, document stores) via simple JSON configuration
- Self‑hosted Docker deployment with zero‑config mode and a cloud‑hosted SaaS option for rapid onboarding
- Secure data handling with end‑to‑end encryption and role‑based access controls for meeting transcripts and generated content
- Upcoming screen‑sharing capability that lets the AI present generated slides or visualizations directly to participants
- API endpoints for retrieving chat history, participant lists, and transcript snapshots to integrate with enterprise collaboration platforms