espnet/espnet

Health v2 · maintenance only

Adoption not part of the score

9941 stars · 2425 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-29, confidence not recorded

ESPnet is an end-to-end speech processing toolkit built on PyTorch covering speech recognition, text-to-speech, speech translation, enhancement, and speaker diarization. It provides recipes, pretrained models via a model zoo, and fine-tuning support for research and production speech tasks.

Use cases

• transcribe speech to text with asr models
• convert text to speech with tts
• translate spoken audio to another language's text
• enhance noisy speech recordings
• diarize speakers in a conversation audio
• fine-tune a pretrained speech model on my own data
• separate overlapping speakers in audio

When to choose

• you need state-of-the-art end-to-end speech models with complete training recipes
• you want pretrained ASR/TTS models from a model zoo with easy inference
• you are doing speech processing research and need reproducible experiments

When to avoid

• you need a lightweight production ASR API without training infrastructure
• you want a simple plug-and-play transcription tool with minimal setup
• your task is outside speech/audio processing

Facets

library · maturity active

speech-recognition tts machine-learning deep-learning audio-processing nlp speech-processing deep-learning machine-learning python windows speech-translation voice-conversion speech-enhancement speaker-diarization singing-voice-synthesis pytorch kaldi asr text-to-speech speech-separation natural-language-processing audio linux macos gpu docker

Member repositories

For agents

markdown · JSON · MCP: product_card(name="espnet/espnet")