Wispr Flow raised $260M for AI voice typing; a solo dev cloned the core in a weekend
Wispr Flow, an AI dictation tool that turns spoken words into typed, cleaned-up text, raised $260 million at close to a $2 billion valuation. It has 2.5 million downloads and is used by 270 Fortune 500 companies, charging $15 per month ($180 per year). A founder building a called ChromePilot added a similar feature called Voice Keyboard: hovering over any text field on any website shows a mic icon, clicking it lets you talk, and the transcribed text drops into that field.
It works across Gmail, Slack, Notion, Reddit, and more. The whole feature took one weekend to build. It runs as a content script that detects input fields and injects a mic button using Shadow DOM so it doesn't break the page's styling.
Clicking record captures audio through the browser's , sends it to for , and inserts the result at the cursor. The total is about 600 lines of plain JavaScript with no frameworks or dependencies. The builder acknowledges Wispr Flow does more, since it works system-wide rather than just in the browser.
Key points
- Wispr Flow: $260M raised, ~$2B valuation, 2.5M downloads, used by 270 Fortune 500 companies, $15/month
- A solo builder replicated the core feature as a add-on in one weekend
- Built with a content script detecting input fields, a Shadow DOM mic button, the , and for
- About 600 lines of plain JavaScript, no frameworks or external dependencies
- Wispr Flow's edge is working system-wide across the OS, not just inside the browser