Resource
The useful question is not just “which meeting tool has MCP?” It’s “which tool gives my assistants durable meeting memory in a form they can actually use?” That is a smaller, more specific category, and the tools in it are not all good at the same thing.
Quick answer
Consider Minutes for local meeting records stored as Markdown and searched through your preferred MCP client. Compare Anarlog too if you want an editable local notepad.
If you want a more polished hosted AI note-taking product with MCP access, Granola AI is a stronger fit. If you want a hosted team assistant with broader integration and admin posture, Fireflies.ai or Otter AI may be the better choice.
Local-first MCP meeting memory
Best when you want local processing, inspectable markdown, and a meeting memory layer that works across MCP, CLI, desktop, SDK, and Claude Code plugin workflows.
Polished AI notepad with MCP
Best when you want a refined AI note-taking experience and also want some MCP access into a hosted product.
Hosted team workflow with MCP
Best when you want a hosted meeting assistant, broader integrations, and MCP access in a team-centered SaaS workflow.
Hosted mainstream meeting assistant with MCP
Best when you want a hosted meeting assistant with centralized transcripts, collaboration, and broader team/admin posture.
A local notepad with CLI and MCP
An editable meeting app backed by local SQLite, with Markdown export and a local CLI/MCP surface. Select local or hosted AI providers separately.
Self-hosted capture infrastructure
A meeting-bot API with live transcripts and MCP for people operating their own capture stack. Evaluate the operational setup before choosing it over a desktop app.
Pick Minutes if your assistants need local, inspectable meeting memory they can query later across MCP, CLI, desktop, and plugin workflows.
Pick Granola AI if you want a polished hosted AI notepad with MCP access into that product. Pick Fireflies.ai or Otter AI if you want a hosted meeting assistant with stronger team and admin posture.
Minutes is not the best fit if the real goal is a managed hosted assistant for a team, with centralized admin, broader SaaS integrations, and a workflow built around one polished product experience.
It fits workflows that need local ownership, open artifacts, and agent readability. If that is not the job, one of the hosted tools in this category may be a better fit.
Next step
The same question from the coding-agent side.
Where the meeting notes come from.
The transcription layer underneath.
The file format an assistant reads best.