Contextli connector · MCP

Ask Claude about anything you said.

The Contextli connector lets Claude, ChatGPT, Cursor and VS Code search and read your own voice notes, and save or label notes when you ask. Sign in once with your Contextli account, then ask for a note the way you would ask a colleague.

Claude

What did I say about the pricing call last Tuesday?

contextli_search_notes · “pricing call”
Voice note · Tue 9:42 · Meetings

Pricing call went well. They want annual billing, a single invoice for the whole team, and a call with their IT lead before the end of the month.

Three follow-ups from Tuesday: switch them to annual billing, send one invoice for the team, and book time with their IT lead this month.

Setup

One address works in every app.

Add this server URL to your AI app. There is no key to copy: the first time the app connects, it opens a Contextli sign-in page.

Server URLhttps://mcp.contextli.com/mcp

Claude

  1. Open Settings, then Connectors.
  2. Choose Add custom connector and name it Contextli.
  3. Paste https://mcp.contextli.com/mcp and select Add.
  4. Select Connect and sign in with your Contextli account.

Works in Claude on the web, the desktop app and mobile once added.

ChatGPT

  1. Open Settings, then Apps and Connectors. If you do not see a Create button, turn on developer mode under Advanced settings.
  2. Create a connector named Contextli.
  3. Use https://mcp.contextli.com/mcp as the MCP server URL, with OAuth.
  4. Sign in with your Contextli account when asked.

Cursor

Add this to ~/.cursor/mcp.json. Sign-in opens in your browser on first use.

{
  "mcpServers": {
    "contextli": {
      "url": "https://mcp.contextli.com/mcp"
    }
  }
}

VS Code

Add this to .vscode/mcp.json, then start the server from the MCP view.

{
  "servers": {
    "contextli": {
      "type": "http",
      "url": "https://mcp.contextli.com/mcp"
    }
  }
}

Claude Code

claude mcp add --transport http contextli https://mcp.contextli.com/mcp
# then run /mcp inside Claude Code and sign in

Want your coding agent to work from voice notes on a schedule? See Loop Engineering.

Tools

Eight tools to find, read, save and label your notes.

Your AI app picks the right tool from what you ask. Six of them only read. Two save a note or add a label, and only when you ask. Each one works on your own account only.

  • contextli_search_notesRead

    Search your voice notes and transcriptions by a keyword or phrase you remember saying. Returns matches newest first, with a snippet around each match.

  • contextli_get_notesRead

    List your notes newest first with their full text, filtered by date range, context or label.

  • contextli_get_noteRead

    Open one note by id: full text, metadata and up to 3 attached images, such as a screenshot you saved with it.

  • contextli_list_contextsRead

    List your contexts (modes such as Email or Journaling) and the filters that actually match your notes, so the assistant never filters on an empty value.

  • contextli_list_labelsRead

    List your labels, so the assistant knows what you tag notes with.

  • contextli_filter_notes_by_labelRead

    Get the notes that carry one or more labels, optionally within a context or date range.

  • contextli_create_noteWrite

    Save a new note from the chat, such as a summary or a plan, in the context you choose and with any labels you want.

  • contextli_add_label_to_noteWrite

    Add one or more labels to a note you already have, so it is easy to find later. Existing labels stay as they are.

Things to ask

  • “Search my Contextli notes for everything I said about the onboarding redesign this month and summarise it.”
  • “List yesterday’s voice notes in my Meetings context and pull out the action items.”
  • “Find the note where I described the new landing page idea and open it with its images.”
  • “Use my latest voice note in the Coding context as the spec and start implementing it.”
  • “Save this plan as a note in my Work context and label it launch.”
  • “Add the label pricing to every note from this week where I talked about pricing.”
Sign-in and privacy

You approve the connection, and you can end it any time.

Standard OAuth sign-in

The app connects through OAuth 2.1 with PKCE. You sign in on a Contextli page with a one-time code sent to your email, or with Google or Apple. The app never sees your password.

Scoped to your account

Every request is checked on our server against the signed-in account. The connector never returns another person's notes, your audio recordings or your payment details.

Only what you ask for

Notes go to the AI app only when you ask it something that needs them, and it saves or labels a note only when you ask it to. Nothing is sent in the background, and the connector is not used to train AI models.

Disconnect in one step

Remove Contextli from your AI app's connector settings to revoke access. Access tokens last one hour and sign-ins expire after 30 days without use.

Full details are in section 5.1 of our Privacy Policy and in our Terms.

Need a hand with the connector?

Email support@contextli.com with the app you use and what happened. Not on Contextli yet? Your notes need to be in Contextli first.