FunASR

FunASR is an industrial speech recognition toolkit for offline, streaming, and edge deployment.

FunASR is an industrial speech recognition toolkit designed for offline, streaming, and edge deployment. It provides tasks, checkpoints, and runtimes separately for audio processing like ASR, VAD, punctuation, and speaker diarization. Users can deploy models via Python AutoModel, CLI, vLLM, or llama.cpp for CPU and edge devices. It supports models such as Fun-ASR-Nano, SenseVoiceSmall, and Paraformer. The toolkit is intended for developers building voice agents, AI agents, and transcription pipelines.

FunASR at a glance

Best for: Software engineer, Researcher
Runs on: Linux, Windows, API, CLI, self-hosted, open source
Pricing: open source
Released: November 2022
Category: AI audio tool
Website: github.com/modelscope/FunASR

Key features

Offline file transcription
Streaming live transcription
Speaker embedding extraction
Voice activity detection
Punctuation restoration
Emotion and audio-event recognition
OpenAI-compatible serving

Use it for

Transcribe audio files with emotion and event tags
Run chunked live transcription streams
Perform speaker-aware file transcription
Deploy speech recognition on edge devices

Alternatives to FunASR

Free ASR — Zero-dependency real-time streaming speech-to-text with no account login required.
AssemblyAI — Industry-leading speech models deliver high accuracy and low latency for voice applications.
Superwhisper — Transcribes speech offline and online with custom AI modes and templates.
Speaches — It acts as an Ollama equivalent for running local TTS and STT models.
sherpa-onnx — Runs complex audio AI tasks offline on local devices and embedded hardware.
Whisper ASR Webservice — Provides a REST API for multiple ASR engines including WhisperX and Faster Whisper.
WeNet — Achieves state-of-the-art results on public speech datasets.
whisper.cpp — It provides high-performance, offline speech recognition on diverse hardware.
Vosk — It enables offline, low-latency speech recognition on small devices.
FireRedAudio — A unified model handling understanding and generation with decoupled continuous representations.
Speech-to-Speech — Fully modular, swappable components with local or hosted execution options.
FluidAudio — Runs local audio AI models on Apple Neural Engine.

All FunASR alternatives