Miro MCP

Miro's own MCP server connects your agent to your Miro boards. It can search and summarize what's on a board, build boards for workshops, draw flowcharts and technical diagrams, and read and resolve comments. It reads a board as SVG and writes changes back the same way, so it can change one frame without touching the rest. What you can do with it - Summarize a board, or cluster a research wall's sticky notes into themes - Build boards for retros, sprint planning and strategy sessions, with frames, sticky notes, shapes, cards, docs and tables - Draw flowcharts, architecture, sequence, UML class and ERD diagrams from a description, your code or a pull request - Create docs, tables, timelines and kanban boards, and add images from public URLs - Read, reply to and resolve comments, and turn the discussion into next steps - Create prototypes from HTML screens, and read the screens of existing ones - Search your boards and create new ones How it connects Miro hosts it at https://mcp.miro.com/. The first time your agent uses it, you sign in to Miro with OAuth and pick the team whose boards it can use. It only sees boards you can already open, and your board permissions still apply. In Gemini CLI it installs as an extension (gemini extensions install https://github.com/miroapp/miro-ai), which also brings Miro's skills. Good to know - It works on every Miro plan, including Free. Each plan has a daily allowance of tool calls; past it, calls are refused until the limit resets at 00:00 UTC. - On Enterprise plans an admin has to turn it on first, under Admin settings, Apps & Integrations, MCP. - It's connected to one team at a time. To use boards in another team, sign in again while viewing a board in that team. - Put the full board URL in your ask so it finds the right board.