MellowMill/Documentation/Core features/Bring your own AI
Core feature

Bring your own AI

Connect the AI tooling you already trust through MCP, then ask questions about notes that carry structure, decisions, and context.

Why structured notes help

Margin gives your notes more than formatting. It can distinguish tasks, decisions, justifications, tags, and links, so a connected AI tool has useful signals about how the work is organized.

That makes it easier to ask grounded questions about what you decided, why you decided it, and what needs to happen next.

Connect through MCP

  1. Choose your AI tool

    Bring the provider or local workflow that fits your work. Claude Code is supported today; ChatGPT, Gemini, and local models are planned.

  2. Connect MellowMill through MCP

    Use the MCP connection supported by your AI client to make the relevant MellowMill workspace available to it.

  3. Ask about the work

    Start with a question that depends on context, such as which decisions are open or which tasks are connected to a project.

Questions with useful context
Which decisions changed the launch scope?

Show the open tasks for the onboarding project.

What reasons did we record for choosing the current flow?

Understand the data boundary

MellowMill remains local-first: your workspace files stay on your machine. When you connect an AI tool, the relevant data is sent to that AI provider so it can answer your request. Review the provider's terms and privacy policy before connecting sensitive work.

You choose which AI tooling to bring into the workflow. MellowMill does not turn that choice into a recurring cloud workspace.

Check the boundary before you connect. Local storage protects your workspace by default; an AI request still follows the data-handling rules of the provider you choose.