Wispr Flow launches a Granola-styled meeting notetaker

Wispr Flow launched its meeting notetaker for Mac on August 5. The product is Granola with a different cap table. It taps system audio, transcribes in real time, and never joins the call as a participant. No recording lands on disk. The transcript appears live. An AI sidebar lets you ask what you missed. Action items and summaries generate automatically. The company calls it a notetaker. The terms of service call it a data pipeline.

The updated terms, pushed live before the launch, define Input broadly: meeting audio, participant information, meeting metadata, speaker labels, meeting transcripts, and other information processed in connection with a meeting. Output includes AI-generated meeting transcripts, summaries, action items, meeting insights, speaker attribution, and other meeting-related content. That language is not boilerplate. It is a permit slip. Wispr Flow now sits on a firehose of organizational conversation. Every standup, every board review, every compensation talk flows through their inference layer.

Granola proved the model. System audio avoids the social friction of a bot avatar. It sidesteps the compliance nightmare of stored recordings. It works on Zoom, Meet, Teams, Webex, Around — anywhere sound hits the sound card. Wispr Flow copied the architecture. They had the distribution. Their dictation app already lives in the menu bar of developers, writers, executives. Adding a notetaker is a single toggle. The install base is the moat.

The valuation math is the tell. $700 million last round. $2 billion rumored next round. Bloomberg reported the talks in May. Dictation alone does not command a two-billion-dollar multiple. Context does. Meeting transcripts are the richest unstructured data source inside a company. They reveal roadmap, strategy, hiring plans, legal exposure, product priorities. An AI assistant that has ingested six months of your leadership team calls does not need prompting. It anticipates. It drafts the memo. It flags the contradiction between what the CTO said Tuesday and what the CEO promised Thursday.

Fireflies, Read AI, Otter, Fathom — they all join as participants. They record. They store. They trigger IT reviews. They generate administrative overhead. Granola and now Wispr Flow refuse that tax. But Granola stayed lean. Wispr Flow raised eighty-one million dollars. Investors expect a platform play. The notetaker is the wedge. The assistant is the product. Kothari has said as much in interviews. The meeting data trains the model that will eventually automate the work the meetings are about.

Privacy policies are fiction until tested. Wispr Flow did not respond to a request for comment on data retention, encryption at rest, subprocessor agreements, or whether meeting transcripts improve the base dictation model. The terms grant broad license. The product processes audio locally — they claim — but the AI features require cloud round trips. Summaries, action items, cross-meeting queries: those run on GPUs somewhere. The audio touches their infrastructure. The transcripts live in their database. The insights belong to them.

Competitors will argue feature parity. Better speaker diarization. Cleaner action-item extraction. Slack integration. Notion sync. Salesforce push. Those are table stakes. The real game is data gravity. The first notetaker a team adopts becomes the system of record. Switching costs compound weekly. Wispr Flow enters with an installed base that trusts its dictation engine. That trust transfers. The menu bar icon is already muscle memory.

The skepticism writes itself. A dictation startup pivots to meetings because the TAM is wider and the data is stickier. The product launches before the security white paper. The terms expand before the marketing site updates. The valuation leaks before the revenue proves. This is the playbook of the 2021-2024 vintage. It worked for Fireflies. It worked for Otter. It may work for Wispr Flow. But the market is saturating. IT buyers are consolidating. The next round at $2 billion requires evidence that meeting data converts to recurring revenue outside the notetaker itself.

Users get a competent tool today. Live transcript. Clean summaries. Query across history. No bot in the gallery view. The Mac app is native, fast, unobtrusive. If you already pay for Wispr Flow, the notetaker is included. That price — zero marginal cost — will win pilots. Pilots become contracts. Contracts become the dataset. The dataset becomes the moat.

Watch the next terms update. Watch for an API. Watch for a "Wispr Assistant" that drafts your Jira tickets from Tuesday's standup. That is the product the investors underwrote. The notetaker is just the collection mechanism.