handy-computer/transcribe.cpp

Health v2 · maintenance only

Flags: young

Adoption not part of the score

1836 stars · 92 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

A C/C++ speech-to-text inference library built on the ggml runtime that runs 16+ ASR model families (Whisper, Parakeet, Canary, Moonshine, Qwen3-ASR, Voxtral, and more) from GGUF files. It supports streaming and batch transcription with Metal, Vulkan, and CUDA GPU backends plus an accelerated CPU path, with quantized models published and WER-verified on Hugging Face.

Use cases

• transcribe audio files to text locally
• run whisper models without python
• streaming speech-to-text in real time
• transcribe multi-speaker audio with diarization
• run ASR on GPU with metal or cuda
• quantized speech recognition on cpu
• multilingual speech transcription offline

When to choose

• you need fast, local speech-to-text inference in C/C++ without Python dependencies
• you want to run many different ASR model families through one unified runtime
• you need streaming transcription or speaker diarization on-device
• you want quantized GGUF models with verified WER parity to reference implementations

When to avoid

• you need to train or fine-tune ASR models rather than run inference
• you want a managed cloud transcription API
• you need TTS or full audio-LLM features beyond transcription and translation

Facets

library · maturity active

speech-recognition llm-inference cli speech-processing machine-learning developer-tools cpp windows cross-platform asr ggml gguf whisper parakeet streaming-transcription diarization metal vulkan cuda gpu macos linux

Member repositories

For agents

markdown · JSON · MCP: product_card(name="handy-computer/transcribe.cpp")