TwinMind is a mobile-first AI app that runs in the background (with user permission) to capture ambient speech, transcribe it on-device, and turn conversations into a personal knowledge graph. The product offers continuous on-device transcription, offline capability (via Ear-2), a Chrome extension for browser-context capture, and a cloud-backed larger model (Ear-3) with an API. TwinMind reports over 30,000 users and about 15,000 monthly active users; roughly 20–30% of users also use the Chrome extension, and the team of 11 plans to hire designers and build a business development team. The company has a Pro subscription at $15/month (larger context window) while keeping a free tier with unlimited hours of transcriptions and on-device recognition. TwinMind emphasizes privacy: audio is deleted on the fly, only transcribed text is stored locally, and the company says it does not train its models on user data. The Ear-3 model supports 140+ languages, has a reported word error rate of 5.26% and speaker diarization error of 3.8%, will be available via API priced at $0.23/hour, and the app can switch between on-device and cloud models as connectivity changes.

Total raised

$6M

Funding rounds

1

Latest round

Seed

Latest activity

Sep 2025