Miro’s MCP server gives an agent access to boards: find them, summarise them, draw on them and act on their comments. It works on every Miro plan, including free teams, and respects existing board permissions.
https://mcp.miro.com/
Claude Sonnet 5
Chat with 60+ AI models on the same workflow — switch to a different model mid-conversation and re-run the same prompt, or use Compare mode to put several side-by-side and balance quality vs. cost.
OAuth 2.1 with dynamic client registration. You sign in to Miro and choose the team the agent can access.
How models use it and what it is built for.
The interesting part is the canvas. Rather than placing shapes one call at a time, the agent reads a board region as SVG with canvas_read_as_svg, and writes with canvas_create_from_svg or canvas_update_from_svg, which applies changes as diffs. A dedicated tool returns the SVG format specification, so the model can learn the layout rules before it draws. Around that sit board search and creation, comments (list, reply, resolve), images, table rows, and prototypes built from HTML screens. In practice that covers turning meeting notes into a flowchart, summarising a retro board, or answering the open comments on a design review.
Typical tools an AI model can call. Exact names vary by version.
Copy any of these into MCP Agent Studio after connecting.
Summarise the Q3 retro board: top themes, and anything with more than three votes.
Turn these onboarding steps into a flowchart on a new board.
List unresolved comments on the checkout redesign board and draft a reply to each.
Add a swimlane for the payments team to the architecture diagram.
This is not a single-model product: you get the same MCP connection with 60+ models (Claude, GPT, Gemini, DeepSeek, open-weight, and more), you can switch mid-conversation, and you can open Compare mode to run the same prompt against multiple models at once. The card above is a suggested starting point for this server — not the only choice.
Default pick for Miro
Claude Sonnet 5
Drawing through SVG diffs needs a model that writes well-formed SVG and keeps coordinates consistent. Sonnet 5 handles that layout reasoning reliably.
Listing tools proves the server is reachable, not that a model can work with it. Evals go further: they read every tool on the server, write a test suite from its real schemas, and run it — code decides pass/fail on the responses (schema conformance, error codes, pagination, result caps) while a scoring model grades plain-English tasks driven through the tools.
Get a pass/fail report per tool with the evidence behind each verdict — and replay the same suite after every schema change. Destructive tools are excluded from the run.
Run evalsOpen MCP Agent Studio with the connection pre-filled. Add your token, pick any of 60+ models, and start chatting — no install required.
Open Agent StudioCommon questions about connecting, scoping and using it safely.
Yes. Miro says it works on all plans, including free teams. Enterprise admins get extra security controls.
Through SVG. It reads existing items with canvas_read_as_svg and writes with canvas_create_from_svg or canvas_update_from_svg, which applies changes as diffs. canvas_get_canvas_composer_skill gives the model the format rules first.
With OAuth 2.1. Clients that support dynamic client registration can connect to https://mcp.miro.com/ and sign in to Miro in the browser.
No. It respects existing board permissions and access controls.
Yes. Standard API rate limits and daily MCP usage limits apply. Miro documents them on its MCP limits page.