Sign in Submit site

FunASR

github.com/modelscope/FunASR AI Speech Transcription Open-source projectOpen source

FunASR supports speech recognition training and inference, streaming, punctuation, diarization, and model serving.

Visit FunASR 12 alternatives
github.com/modelscope/FunASR
FunASR homepage screenshot

About FunASR

FunASR is an open-source speech recognition toolkit hosted on GitHub. It covers model training and inference, streaming speech recognition, voice activity detection (VAD), punctuation, and speaker diarization pipelines. It can also serve models through OpenAI-compatible and MCP interfaces.

The repository lists several model options. Fun-ASR-Nano covers Chinese, English, Japanese, and Chinese dialect groups and regional accents. Fun-ASR-MLT-Nano-2512 is a separate checkpoint covering 31 languages. SenseVoiceSmall is a CPU-first option that provides five-language ASR along with emotion and audio-event tags. Language coverage depends on the specific checkpoint, so the models should be chosen separately.

Limits: GPU use requires PyTorch and torchaudio wheels that match the user's NVIDIA driver. Speaker indices from the pipeline are local to each recording and do not identify known people. The speaker outputs come from the pipeline, not from the SenseVoiceSmall checkpoint itself. A browser-based Colab quickstart is available for transcribing audio without local setup, and uploaded audio is processed in that Colab environment.

Key features

  • Streaming speech recognition and voice activity detection (VAD)
  • Speaker diarization pipelines that assign anonymous speaker indices to segments
  • Punctuation support in transcription pipelines
  • OpenAI-compatible and MCP model serving
  • Training and inference tooling in one toolkit
  • SenseVoiceSmall model with emotion and audio-event tags
Best for: Developers and researchers who want open-source speech recognition models and tools they can run, customize, or deploy themselves.
Good for
  • Open source, with the repository available on GitHub
  • Multiple checkpoints for different language sets, including a CPU-first option
  • A browser-based Colab quickstart is available for trying transcription without local setup
Watch out for
  • Language coverage is specific to each checkpoint
  • GPU use requires PyTorch and torchaudio wheels that match the user's NVIDIA driver
  • Speaker indices are local to a recording and are not known-person identities

FunASR at a glance

Type
Open-source project
Pricing
Not confirmed. Check the site
Open source
Yes · Source code
Language
English
Last checked
Oct 6, 2026 · Online when checked
Listed
Oct 5, 2026
open sourcechinesespeech recognitionmultilingual

FunASR alternatives

Tools and sites you could use instead, ranked by how closely they match.

All 12 alternatives

Popular in AI Speech Transcription

Well-known listings from the same topic.

All 43
Transgatetransgate.aiWavewave.coRev.airev.aiGood Tapegoodtape.ioTalkNotestalknotes.ioSpeech to Notespeechtonote.com
Related topicsAI Business ToolsAI Image GeneratorsAI Infrastructure & MLOpsAI Agents & AutomationAI Data & AnalyticsAI Chatbots & AssistantsAI Voice & AudioAI Writing Tools

FunASR FAQ

Do I need a local setup to try FunASR?
No. The repository links a Colab quickstart that can transcribe a public sample or user-uploaded audio in a browser. Uploaded audio is processed in that Colab environment.
Does FunASR identify who is speaking?
No. The speaker indices assigned by the pipeline are local to each recording and are not known-person identities.
Which languages does FunASR support?
It depends on the checkpoint. Fun-ASR-Nano covers Chinese, English, Japanese, and Chinese dialect groups. Fun-ASR-MLT-Nano-2512 covers 31 languages. SenseVoiceSmall covers five languages.
Is something wrong or out of date?

Tell us and an editor will check it.

Own this website?

Claim the listing and suggest updates yourself. An editor checks every claim first.

Claim this listing

Listing maintained by Chris McClure and the WebIndexly editors · Updated Oct 12, 2026 · How we rank