Source: https://seerepo.com/r/huggingface/speech-to-speech/built
Markdown source: https://seerepo.com/r/huggingface/speech-to-speech/built.md
Description: speech-to-speech is a Python repo built with FastAPI, anyio, and uvicorn. It uses AI through LiveKit, torch, transformers, mlx-audio, librosa, and 7 more.

# What is speech-to-speech built with?

speech-to-speech is a Python repo built with FastAPI, anyio, and uvicorn. It uses AI through LiveKit, torch, transformers, mlx-audio, librosa, and 7 more, makes its content with pillow, validates its data with Pydantic and jsonschema, and uses Hugging Face. It has 106,156 written lines, with 155 test files.

Read on November 9, 2026, at commit 327a8c1.

Repository: https://github.com/huggingface/speech-to-speech

## Frontend

### Interface

- **[Rich](https://seerepo.com/uses/rich)**: 6 files

## Backend

### The base

- **[FastAPI](https://seerepo.com/uses/fastapi)**: 3 files

### AI

- **LiveKit**
- **[torch](https://seerepo.com/uses/torch)**: 23 files
- **[transformers](https://seerepo.com/uses/transformers)**: 11 files
- **mlx-audio**: 5 files
- **[librosa](https://seerepo.com/uses/librosa)**: 4 files
- **[mlx](https://seerepo.com/uses/mlx)**: 4 files
- **[nltk](https://seerepo.com/uses/nltk)**: 3 files
- **faster-qwen3-tts**: 1 file
- **mlx-lm**: 1 file
- **[onnxruntime](https://seerepo.com/uses/onnxruntime)**: 1 file
- **[torchaudio](https://seerepo.com/uses/torchaudio)**: 1 file

### Services

- **[OpenAI](https://seerepo.com/uses/openai)**: 4 files · [src/speech_to_speech/LLM/base_openai_compatible_language_model.py, line 17](https://github.com/huggingface/speech-to-speech/blob/327a8c190d4126f8f7e06287b1086078835a5a4a/src/speech_to_speech/LLM/base_openai_compatible_language_model.py#L17)
- **[Hugging Face](https://seerepo.com/uses/hugging-face)**: 3 files · [src/speech_to_speech/STT/nemo_asr_handler.py, line 120](https://github.com/huggingface/speech-to-speech/blob/327a8c190d4126f8f7e06287b1086078835a5a4a/src/speech_to_speech/STT/nemo_asr_handler.py#L120)

## Support

### The base

- **[anyio](https://seerepo.com/uses/anyio)**: 1 file
- **[uvicorn](https://seerepo.com/uses/uvicorn)**: 1 file

### Content

- **[pillow](https://seerepo.com/uses/pillow)**: 1 file

### Validation

- **[Pydantic](https://seerepo.com/uses/pydantic)**: 14 files
- **[jsonschema](https://seerepo.com/uses/jsonschema)**: 1 file

### Other

- **certifi**
- **miniaudio**
- **NumPy**: 41 files
- **SciPy**: 9 files
- **httpx**: 5 files
- **soxr**: 4 files
- **soundfile**: 2 files
- **websockets**: 2 files
- **pyyaml**: 1 file
- **sounddevice**: 1 file

## Other repos with a similar stack

- [mudler/LocalAI](https://seerepo.com/r/mudler/LocalAI/built): FastAPI, OpenAI, Hugging Face, anyio, uvicorn, torch, transformers, mlx-audio, librosa, mlx, faster-qwen3-tts, mlx-lm, onnxruntime, torchaudio, pillow · [Compare](https://seerepo.com/compare/huggingface/speech-to-speech/vs/mudler/LocalAI)
- [bojieli/ai-agent-book](https://seerepo.com/r/bojieli/ai-agent-book/built): FastAPI, OpenAI, Hugging Face, anyio, uvicorn, Rich, torch, transformers, mlx-audio, librosa, nltk, torchaudio, pillow, Pydantic · [Compare](https://seerepo.com/compare/bojieli/ai-agent-book/vs/huggingface/speech-to-speech)
- [debpalash/VoiceStudio](https://seerepo.com/r/debpalash/VoiceStudio/built): FastAPI, OpenAI, Hugging Face, uvicorn, Rich, torch, transformers, librosa, onnxruntime, torchaudio, pillow, Pydantic · [Compare](https://seerepo.com/compare/debpalash/VoiceStudio/vs/huggingface/speech-to-speech)
- [datawhalechina/hello-agents](https://seerepo.com/r/datawhalechina/hello-agents/built): FastAPI, OpenAI, Hugging Face, anyio, uvicorn, Rich, torch, transformers, onnxruntime, pillow, Pydantic, jsonschema · [Compare](https://seerepo.com/compare/datawhalechina/hello-agents/vs/huggingface/speech-to-speech)
- [ray-project/ray](https://seerepo.com/r/ray-project/ray/built): FastAPI, Hugging Face, anyio, uvicorn, Rich, torch, transformers, nltk, onnxruntime, pillow, Pydantic, jsonschema · [Compare](https://seerepo.com/compare/huggingface/speech-to-speech/vs/ray-project/ray)
- [gradio-app/gradio](https://seerepo.com/r/gradio-app/gradio/built): FastAPI, OpenAI, Hugging Face, anyio, uvicorn, torch, transformers, librosa, nltk, torchaudio, pillow, Pydantic · [Compare](https://seerepo.com/compare/gradio-app/gradio/vs/huggingface/speech-to-speech)

## Frequently asked questions

### What language is speech-to-speech written in?

speech-to-speech is written in Python, with 106,156 written lines.

### What framework does speech-to-speech use?

speech-to-speech is built with FastAPI, anyio, and uvicorn.

### Does speech-to-speech use AI?

Yes. speech-to-speech uses AI through LiveKit, torch, transformers, mlx-audio, librosa, mlx, nltk, faster-qwen3-tts, mlx-lm, onnxruntime, torchaudio, and OpenAI.

### What services does speech-to-speech use?

speech-to-speech uses Hugging Face.

The proof: [src/speech_to_speech/STT/nemo_asr_handler.py, line 120](https://github.com/huggingface/speech-to-speech/blob/327a8c190d4126f8f7e06287b1086078835a5a4a/src/speech_to_speech/STT/nemo_asr_handler.py#L120)

### Does speech-to-speech have tests?

Yes. speech-to-speech has 155 test files.

### Which repos have a stack like speech-to-speech's?

The ones that share most of its stack: mudler/LocalAI, bojieli/ai-agent-book, and debpalash/VoiceStudio.

## What is your repo built with?

Any public repo, with no account: https://seerepo.com/r/owner/repo/built, with the owner and name it has on GitHub.
