Prefer to manage Kolab from inside Claude or another AI assistant? Kolab runs a remote MCP (Model Context Protocol) server, so you can connect an MCP-compatible client—like Claude Desktop or claude.ai—directly to your Kolab spaces and drive them without opening the app.
Connecting
Add Kolab's MCP server as a connector or integration in your AI client, using this address:
https://mcp.getkolab.ai/mcp
There's no separate API key or token to generate. Connecting kicks off a standard login flow using your existing Kolab account—however you normally sign in, whether that's Google, LinkedIn, or an email link.
What It Can Do
Once connected, your AI assistant can:
- Manage spaces — create, update, copy, delete, and adjust access
- Build and edit canvas elements — stickies, shapes, text, and more
- Activate session-aid tools — timers, prompts, card decks, embeds, sticky banks
- Set up agendas and breakout rooms
- Read and search transcripts from past sessions, including AI summaries and speaking-time analytics
- Pull space and team insights
- Hand off bigger build requests to Kolab's own AI facilitator using the
facilitatecapability, for asks too involved to build tool call by tool call
A note on newer features: anything still in limited beta in the Kolab app (like Interactives) stays gated the same way through MCP—connecting doesn't unlock early access to features your account can't already use.
This is a practical way to drive Kolab from outside the app—ask Claude to spin up a space, prep an agenda, or pull a summary of your team's activity, all without opening Kolab directly.
What It Can Access
Access through MCP is scoped to exactly what your logged-in account can already reach in Kolab. Spaces and teams you don't have access to in the app aren't accessible through MCP either—connecting doesn't grant any new permissions, it just gives your assistant a way to use the ones you already have.
Related Guides
- Team Administration - Manage the teams and access your MCP connection can reach
- Creating a Space - Set up spaces your AI assistant can build on
- Tools vs. Canvas Elements - What "tools" and "elements" mean in Kolab
- Live Transcription - What's in a transcript your assistant can read