Wispr Flow is best known as a dictation app, but it also has a “Notetaker” feature that records and summarizes meetings. Recently, Wispr Flow released an official MCP server, which means AI tools like Claude and ChatGPT can now directly search and reference the meeting notes stored in Wispr Flow.
In this article, I’ll cover what you can do with Wispr Flow’s MCP integration, what the benefits are, and the limitations worth knowing before you use it.
What Is MCP?
MCP (Model Context Protocol) is a common standard that lets AI apps connect to external services and data. Once you connect an MCP-compatible service, the AI can go and look inside that service.
Until now, you had to copy a meeting transcript yourself and paste it into an AI chat. With MCP connected, the AI can find the relevant meeting and read it on its own.
What You Can Do with Wispr Flow’s MCP

With Wispr Flow’s MCP, the AI can read the following data:
| What it can read | Details |
|---|---|
| Meetings (Notetaker) | Summaries, notes, transcripts, participants, tasks |
| Calendar | Events, linked recordings, upcoming meetings, recurring meetings |
| Scratchpad | Notes inside Wispr Flow |
| Shared notes | Notes shared with you via a share link |
It can also match the corresponding event from a calendar or web meeting link. If you’ve corrected speaker names in Notetaker or linked them to contacts, the AI receives those names instead of “Speaker 1.”
Example Prompts
For example, you can ask things like:
- What did we decide in yesterday’s pricing meeting?
- Summarize this month’s calls with Acme and list any open action items
- Help me prepare for my next meeting
- Find my scratchpad notes on the hiring plan
You can also use it in ways like these:
- Look back over several weekly meetings and pull out the items that are still undecided
- Have it read the summary of a specific meeting and draft a follow-up email using another connector
- Review your last conversation with a customer, then combine it with your next scheduled meeting to create prep notes
Supported AI Tools
The same server URL (https://api.wisprflow.ai/connect/mcp) works with multiple clients.
| Tool | How to connect |
|---|---|
| Claude | Add from the connector directory |
| ChatGPT | Add as a plugin |
| Gemini | Enable in the Gemini app |
| Cursor | Install “Wispr Flow” from the plugin marketplace |
| Grok Build | Install “Wispr Flow” from the plugin marketplace |
| Claude Code | claude mcp add --transport http wispr-flow https://api.wisprflow.ai/connect/mcp |
| Other MCP clients | Add the URL above as a remote (HTTP) MCP server |
Wispr Flow’s settings screen (Settings → MCP) also has “Add to ○○” buttons for Claude, ChatGPT, Gemini, and Cursor.

Benefits of Wispr Flow × MCP
-
No copy and paste. The AI finds the meeting and reads the summary, notes, and transcript itself
-
It's read-only, so there's no risk of the AI modifying or deleting your meeting data
-
No API keys or tokens to manage. Just log in to Wispr Flow in your browser and grant permission
-
One URL works across multiple AI tools, including Claude, ChatGPT, and Cursor
-
You can query across not just meetings but also your calendar and scratchpad
-
It can read the data in your account without the Wispr Flow app having to stay open
-
Permission is granted at the account level, with no way to limit it by meeting or time period
-
Search matches titles, summaries, speaker names, and similar fields. You can't find a meeting by a word that appears only in the transcript body
-
Search results are ordered by most recently updated, not by meeting date
-
It can't read recordings from other note-taking tools (such as Granola or Otter)
How to Connect
:::steps
- Check the prerequisites
Make sure your Wispr Flow account (Mac or Windows) has Notetaker and MCP available. If you’re adding a custom connector in Claude or a similar tool, the AI tool itself needs to be on a paid plan.
- Open Wispr Flow’s settings
Go to Wispr Flow → Settings → MCP.
- Click “Add to ○○” for the AI you want to use
Choose from the Claude, ChatGPT, Gemini, or Cursor buttons. The AI tool’s settings page opens in your browser, where you continue the connection.
- Sign in to Wispr Flow and grant permission
Sign in in your browser and click “Allow” on the authorization screen. If you sign in with Google or enterprise SSO, you need to finish within 5 minutes. Signing in with an email address and password is reported to be unable to complete the connector authorization, so use Google or SSO.
- Enable the connector in chat and test it
In Claude, click ”+” in the chat box → “Connectors” → turn on “Wispr Flow.” Then ask about a recent meeting. If the actual contents of the meeting come back, the connection is working.
Things to Know Before You Use It
Permission Covers the Whole Account
Once you authorize it, the AI can read every meeting in your account, including past ones. There’s no per-meeting consent and no way to specify a time range. Personal and work accounts are separate, so check in your browser which one you’re granting access to.
What It Can’t Read
Because it’s read-only, the AI can’t create, edit, share, or delete anything. It also can’t change sharing settings. In addition, it can’t read your dictation history or the chat history inside Notetaker.
Limitations Such as Cloud Sync and HIPAA
- Using Notetaker meetings and transcripts may require Cloud Sync (per the README requirements).
- MCP features aren’t available on HIPAA-covered accounts.
- For organizations (Enterprise), an admin must have enabled MCP.
- Transcripts have a retention setting, and deleted transcripts can no longer be read through MCP either. The default retention period for recording data is 7 days from when the upload completes.
How Shared Notes Are Handled
For notes shared with you, only the title and summary are searchable. You can read the transcript only if the owner invited you by email and the owner’s organization allows transcript sharing.
Summary
Wispr Flow’s MCP integration isn’t about “handing” your meeting records to the AI. It’s a way to let the AI go and find them itself. It’s read-only and needs no API key, which makes it easy to try.
If you already use Wispr Flow’s Notetaker, connecting it to the Claude or ChatGPT you already use can make reviewing meetings and preparing for the next one much easier.
