k2-fsa /sherpa-onnx
Speech-to-text, text-to-speech, speaker diarization, speech enhancement, source separation, and VAD using next-gen Kaldi with onnxruntime without Internet connection. Support embedded systems, Android, iOS, HarmonyOS, Raspberry Pi, RISC-V, RK NPU, Axera NPU, Ascend NPU, x86_64 servers, websocket server/client, support 12 programming languages
★ 15,968⑂ 1,878👁 119⑃ 2,396● 1,659▤ 195
+21
Daily
15,968
stars
1,878
forks
119
watchers
2,396
prs
1,659
issues
195
releases
README
Supported functions |Speech recognition| [Speech synthesis][tts-url] | [Source separation][ss-url] | |------------------|------------------|-------------------| | ✔️ | ✔️ | ✔️ | |Speaker identification| [Speaker diarization][sd-url] | Speaker verification | |----------------------|-------------------- |------------------------| | ✔️ | ✔️ | ✔️ | | [Spoken Language identification][slid-url] | [Audio tagging][at-url] | [Voice activity detection][vad-url] | |--------------------------------|---------------|--------------------------| | ✔️ | ✔️ | ✔️ | | [Keyword spotting][kws-url] | [Add punctuation][punct-url] | [Speech enhancement][se-url] | |------------------|-----------------|--------------------| | ✔️ | ✔️ | ✔️ | Supported platforms |Architecture| Android | iOS | Windows | macOS | linux | HarmonyOS | |------------|---------|---------|------------|-------|-------|-----------| | x64 | ✔️ | | ✔️ | ✔️ | ✔️ | ✔️ | | x86 | ✔️ | | ✔️ | | | | | arm64 | ✔️ | ✔️ | ✔️ | ✔️ | ✔️ | ✔️ | | arm32 | ✔️ | | | | ✔️ | ✔️ | | riscv64 | | | | | ✔️ | | Supported programming languages | 1. C++ | 2. C | 3. Python | 4. JavaScript | |--------|-------|-----------|---------------| | ✔️ | ✔️ | ✔️ | ✔️ | |5. Java | 6. C | 7. Kotlin | 8. Swift | |--------|-------|-----------|----------| | ✔️ | ✔️ | ✔️ | ✔️ | | 9. Go | 10. Dart | 11. Rust | 12. Pascal | |-------|----------|----------|------------| | ✔️ | ✔️ | ✔️ | ✔️ | It also supports WebAssembly. Supported frameworks [Flutter][flutter-url] |Architecture| Android | iOS | Windows | macOS | Linux | Web | |------------|---------|---------|---------|-------|-------|-----| | x64 | ✔️ | | ✔️ | ✔️ | ✔️ | ✔️ | | x86 | ✔️ | | | | | ✔️ | | arm64 | ✔️ | ✔️ | ✔️ | ✔️ | ✔️ | ✔️ | | arm32 | ✔️ | | | | | ✔️ | - Pre-built Flutter demo apps: [flutter releases][flutter-releases] [Tauri][tauri-url] |Architecture| Android | iOS | Windows | macOS | Linux | |------------|---------|---------|---------|-------|-------| | x64 | ✔️ | | ✔️ | ✔️ | ✔️ | | x86 | ✔️ | | | | | | arm64 | ✔️ | ✔️ | ✔️ | ✔️ | ✔️ | | arm32 | ✔️ | | | | | - Pre-built Tauri demo apps: [tauri releases][tauri-releases] [flutter-url]: https://flutter.dev [tauri-
Top issues
#1193Questions related to MeloTTS 66 comments · 2024-07-31#2120OpenHamrony 系统上运行项目SherpaOnnxTts 报createCircularBuffer未导出49 comments · 2025-04-14#569Installing multiple voices 35 comments · 2024-02-01#920[Need help] How to realize Syllable-level Voice Recognition with sherpa-onnx Open Vocabulary Keyword Spotting30 comments · 2024-05-25#1081TTS bug28 comments · 2024-07-06#633Whisper onnxruntime exception on Android27 comments · 2024-03-04#176Installing ONNX for Windows gives error for cmake -DCMAKE_BUILD_TYPE=Release ..27 comments · 2023-06-13#2515Rockchip RK3588上npu推理耗时比cpu长23 comments · 2025-08-21
Discussions
#4191OpenValet: sherpa-onnx KWS + Parakeet in an open-source macOS voice agent1 comments · 2026-11-12#4166Murmur (murmurtts.com): Kokoro on Intel Macs with sherpa-onnx0 comments · 2026-11-06#3968Piper vs Kokoro RTF on a 2-core ARM CPU, and what CPU contention costs (not a sherpa-onnx benchmark)2 comments · 2026-09-20#3939License and redistribution of two speaker diarization ONNX models1 comments · 2026-09-09#3903sherpa-onnx + sensevoice android voice input keyboard1 comments · 2026-08-28#3900Sherpa Transcript – 100% offline Android speech-to-text with live speaker diarization (Kroko Zipformer + ERes2Net)0 comments · 2026-08-26#3890altgo:把 sherpa-onnx 内嵌进 Tauri 桌面端做本地语音输入的应用0 comments · 2026-08-24#3834Procedure for hosting sherpa ONNX on Vivante NPU1 comments · 2026-08-03
Latest release
Forced alignment (fa) models fa-models 2026-11-11Acoustic models and pronunciation dictionaries for forced alignment with sherpa-onnx — English (english_mfa v2.2.1 / v3.0.0 / v3.1.0) and Chinese (aishell3_pinyin / biaobei_pinyin / large_pinyin). Models, sources, package contents and usage are documented in scripts/mfa/README.md.