Connect your AI assistant
Let Claude, OpenAI Codex, Cursor, VS Code or another MCP assistant search and read your notes - and, if you allow it, save new ones.
NoteClearly works with AI assistants that support MCP (the Model Context Protocol). Once connected, you can ask your assistant things like “What did I say about the budget last week?” or “Save this as a note”, and it uses your notes to answer.
Set it up
- In NoteClearly, open Settings → AI assistant.
- Choose what the assistant may do:
- Read only (recommended) - search and read your notes, outputs, tags and action items.
- Read and write - also save new notes, tick off and edit action items, tag notes and create tags.
- Click Create token. Copy it now - it’s shown only this once.
- Follow the instructions on the page for your assistant. The token and the server URL are already filled in.
Claude (web and desktop)
In Claude, go to Customize → Connectors → Add custom connector. Enter the MCP server URL from the NoteClearly page, set authentication to No sign-in, and add the request header exactly as shown there (the copy button copies it).
Claude Code
Run the command shown on the page in a terminal.
OpenAI Codex
Add the lines shown on the page to ~/.codex/config.toml.
Cursor, VS Code and others
Add the JSON shown on the page to your client’s MCP settings.
What it can and can’t do
Even with Read and write, an assistant can never edit, delete or share a note, or rename or delete tags. It searches by words, not by meaning, and it doesn’t use your AI credits.
Privacy
Whatever the assistant reads goes to the AI service it runs on - Anthropic, OpenAI and so on - under that service’s terms. Connect only assistants you trust with your notes.
Stop it
Revoke the token under Settings → API Tokens and the assistant loses access at once.
Updated: