k2-fsa/sherpa-ncnn

Real-time speech recognition and voice activity detection (VAD) using next-gen Kaldi with ncnn without Internet connection. Support iOS, Android, Linux, macOS, Windows, Raspberry Pi, VisionFive2, LicheePi4A etc.

64
/ 100
Established

Built on ncnn for lightweight neural network inference, it enables streaming ASR, text-to-speech synthesis, and VAD entirely on-device without external dependencies like PyTorch. Provides bindings across eight programming languages (C++, Python, JavaScript, Go, Swift, Kotlin, C#, WebAssembly) and compiles statically for minimal system requirements, making it suitable for embedded and resource-constrained environments from Raspberry Pi to RISC-V platforms.

1,648 stars and 9,383 monthly downloads. Available on PyPI.

Maintenance 6 / 25
Adoption 19 / 25
Maturity 18 / 25
Community 21 / 25

How are scores calculated?

Stars

1,648

Forks

210

Language

C++

License

Apache-2.0

Last pushed

Oct 20, 2025

Monthly downloads

9,383

Commits (30d)

0

Dependencies

1

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/voice-ai/k2-fsa/sherpa-ncnn"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.