Small streaming speech-recognition models for phones, browsers and low-power devices
Moonshine is listed under Speech & Transcription Engines, in the AI Models & Local Execution category on Flocci AI Tools. Small streaming speech-recognition models for phones, browsers and low-power devices. It is free, with no paid plan attached, and it lives at github.com.
Moonshine is listed as fully free — there is no paid tier attached to it in the catalog. That makes it one of the 397 entries on Flocci AI Tools with no upgrade path built in.
Moonshine needs to be installed or self-hosted — it doesn't run in a browser tab. It is available on the command line, Mac, Windows, Linux, iOS, Android, the web and self-hosting.
Moonshine does 3 things the catalog singles out: Small streaming speech-recognition models for phones, browsers and low-power devices; low latency for live voice interfaces; free open-source weights and runtime.
NVIDIA Canary-Qwen 2.5B is the closest free alternative: it sits in the same Speech & Transcription Engines sub-category and is free. NVIDIA Parakeet, sherpa-onnx and faster-whisper also start free. The full list is on the alternatives page.
See the full list →