# ABexit/ASR-LLM-TTS

This is a speech interaction system built on an open-source model, integrating ASR, LLM, and TTS in sequence. The ASR model is SenceVoice, the LLM models are QWen2.5-0.5B/1.5B, and there are three TTS models: CosyVoice, Edge-TTS, and pyttsx3

Repository: https://github.com/ABexit/ASR-LLM-TTS
Canonical: https://ross.abutalabs.com/products/asr-llm-tts
Language: Python
License: Apache-2.0
License Family: permissive
Last push: 2026-06-03T09:58:27+00:00

## Health v2 (maintenance only)
Score: 60/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 85, release rhythm 35, longevity 47
- inputs: {"age_days": 659, "days_push": 91, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1271, forks 206 (observed 2026-08-28T04:04:11.943315+00:00)

## What it is
An open-source Python speech interaction pipeline that chains SenseVoice ASR, Qwen2.5 LLMs, and TTS engines (CosyVoice, Edge-TTS, pyttsx3) for real-time voice conversations. It adds features like wake-word detection, speaker/voiceprint recognition with CAM++, conversation history memory, interruption handling, and multimodal audio/video input via Qwen2-VL.

## Use cases
- build a local voice assistant with speech recognition and tts
- real-time speech-to-speech conversation with an llm
- custom wake word detection for a voice assistant
- speaker identification for voice chat
- multimodal voice assistant that understands images and video
- interruptible real-time voice chat on gpu

## When to choose
- you want a fully open-source, locally runnable ASR-LLM-TTS voice pipeline in Python
- you need Chinese-focused speech interaction with wake words and speaker recognition
- you want to experiment with different TTS engines or multimodal (audio/video) input

## When to avoid
- you need a production-grade, scalable voice service rather than a research/demo codebase
- you have no GPU or cannot download models from HuggingFace/ModelScope
- you need English-first voice assistant tooling with polished packaging

## Facets
- artifact type: application
- maturity: active
- function: speech-recognition, tts, llm-inference, audio-processing, chatbot
- domain: artificial-intelligence, speech-processing, large-language-models, chatbots
- platform: python, cross-platform
- tags: asr, voice-assistant, sensevoice, qwen, cosyvoice, edge-tts, voice-activity-detection, speaker-recognition, wake-word, real-time, gpu

## Member repositories
- ABexit/ASR-LLM-TTS (main) score 60

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:11.943315+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T05:03:19.963878+00:00, confidence not recorded.
  - readme: https://github.com/ABexit/ASR-LLM-TTS (fetched 2026-08-28T04:04:11.943315+00:00, sha b079e99190de)
- Data as of 2026-08-30T08:39:29.467469+00:00.