We started out building for healthcare—first assisted living, then SNFs, then outpatient clinics—but the idea that stuck wasn’t the vertical. It was voice. We watched elderly patients use dictation to stay connected, and doctors reclaim hours with AI scribes. That’s when we realized dictation shouldn’t be limited to healthcare. If it were fast, accurate, and personal enough, everyone could benefit.

Willow is not basic speech-to-text. It runs in the cloud for low-latency, supports technical terms and custom dictionaries, and learns your formatting and syntax preferences over time. Our users say it feels like it “just works,” even if they’ve never used dictation before. Some things we’ve already built:

1. Real-time dictation with extremely fast output

2. Context-aware text generation

3. Automatic formatting and structure detection

Privacy is our absolute top priority - we do not store or retain voice data by default.

Here are a few simple demos:

1. https://youtu.be/yFxH5HY-72Y?si=krI9WNFKBQE1Hvph

2. https://youtu.be/l2zlECbQQcU?si=fJLX-oYrflOI02kj

3. https://youtu.be/yKTIVCZbwHY?si=ohRMpuAUS8eqceAQ

We’d love feedback from the HN community, technical questions, product ideas, and more! You can try it at https://willowvoice.com. Happy to answer any questions in the comments!