achatbot

achatbot factory, create chat bots with vad,turn, asr, llm(tools)/mllm/audio-llm/omni-llm, tts, avatar, ocr, detect object etc..

• achatbot-go (main/sub agent)

Features

• support Model Context Protocol(MCP) and Agent2Agent (A2A) Protocol
• features

Design

apipeline design

• ⭐️ pipeline design ⭐️

achatbot design

• project structure

Install

[!NOTE] > python --version >=3.10 <=3.12 with asyncio-task

if some other nested loop code with achatbot lib, you need to add the following code:

import nest_asyncio

nest_asyncio.apply()

Tip

use uv + pip to run, install the required dependencies fastly, e.g.: uv pip install achatbot > uv pip install "achatbot[fastapi_bot_server]"

pypi

python3 -m venv .venv_achatbot
source .venv_achatbot/bin/activate
pip install achatbot
# optional-dependencies e.g.
pip install "achatbot[fastapi_bot_server]"

local

git clone --recursive https://github.com/ai-bot-pro/chat-bot.git
cd chat-bot
python3 -m venv .venv_achatbot
source .venv_achatbot/bin/activate
bash scripts/pypi_achatbot.sh dev
# optional-dependencies e.g.
pip install "dist/achatbot-{$version}-py3-none-any.whl[fastapi_bot_server]"

run local lite avatar chat bot

# install dependencies (replace $version) (if use cpu(default) install lite_avatar)
pip install "dist/achatbot-{$version}-py3-none-any.whl[fastapi_bot_server,livekit,livekit-api,daily,agora,silero_vad_analyzer,sense_voice_asr,openai_llm_processor,google_llm_processor,litellm_processor,together_ai,tts_edge,lite_avatar]"
# install dependencies (replace $version) (if use gpu(cuda) install lite_avatar_gpu)
pip install "dist/achatbot-{$version}-py3-none-any.whl[fastapi_bot_server,livekit,livekit-api,daily,agora,silero_vad_analyzer,sense_voice_asr,openai_llm_processor,google_llm_processor,litellm_processor,together_ai,tts_edge,lite_avatar_gpu]"
# download model weights
huggingface-cli download weege007/liteavatar --local-dir ./models/weege007/liteavatar
huggingface-cli download FunAudioLLM/SenseVoiceSmall --local-dir ./models/FunAudioLLM/SenseVoiceSmall
# run local lite-avatar chat bot
python -m src.cmd.bots.main -f config/bots/daily_liteavatar_echo_bot.json
python -m src.cmd.bots.main -f config/bots/daily_liteavatar_chat_bot.json

More details: #161

run local lam_audio2expression avatar chat bot

download model weights

download model weights from huggingface and oss to models folder for local inference, use docker volume to mount models folder

wget https://virutalbuy-public.oss-cn-hangzhou.aliyuncs.com/share/aigc3d/data/LAM/LAM_audio2exp_streaming.tar -P ./models/LAM_audio2exp/
tar -xzvf ./models/LAM_audio2exp/LAM_audio2exp_streaming.tar -C ./models/LAM_audio2exp && rm ./models/LAM_audio2exp/LAM_audio2exp_streaming.tar
git clone --depth 1 https://www.modelscope.cn/AI-ModelScope/wav2vec2-base-960h.git ./models/facebook/wav2vec2-base-960h
huggingface-cli download FunAudioLLM/SenseVoiceSmall  --local-dir ./models/FunAudioLLM/SenseVoiceSmall

run local lam_audio2expression avatar chat bot

2 ways:

• run local lam_audio2expression avatar chat bot (NOTE: don't support MacOS ARM64 architecture)

# install dependencies (replace $version)
pip install "dist/achatbot-{$version}-py3-none-any.whl[fastapi_bot_server,silero_vad_analyzer,sense_voice_asr,openai_llm_processor,google_llm_processor,litellm_processor,together_ai,tts_edge,lam_audio2expression_avatar]"
# if use MacOS ARM64, need pip install tensorflow==2.13.0
# NOTE: python_version < '3.12', spleeter==2.4.2 don't support MacOS ARM64 arch
pip install spleeter==2.4.2
pip install typing_extensions==4.14.0 aiortc==1.13.0 transformers==4.36.2 protobuf==5.29.4

# run http signaling service + webrtc + websocket local lam_audio2expression-avatar chat bot
python -m src.cmd.webrtc_websocket.fastapi_ws_signaling_bot_serve -f config/bots/small_webrtc_fastapi_websocket_avatar_echo_bot.json
python -m src.cmd.webrtc_websocket.fastapi_ws_signaling_bot_serve -f config/bots/small_webrtc_fastapi_websocket_avatar_chat_bot.json
# run http signaling service + webrtc + websocket voice avatar agent web ui
cd ui/webrtc_websocket/lam_audio2expression_avatar_ts && npm install && npm run dev
# run websocket signaling service + webrtc + websocket local lam_audio2expression-avatar chat bot
python -m src.cmd.webrtc_websocket.fastapi_ws_signaling_bot_serve_v2 -f config/bots/small_webrtc_fastapi_websocket_avatar_echo_bot.json
python -m src.cmd.webrtc_websocket.fastapi_ws_signaling_bot_serve_v2 -f config/bots/small_webrtc_fastapi_websocket_avatar_chat_bot.json

• run local lam_audio2expression avatar chat bot with docker container

cd deploy/docker

# build base img
make docker_cpu_debian_img
# build runnable img
# install lam_audio2expression_avatar dependency, need achatbot:base img
make docker_cpu_debian_lam_audio2expression_avatar_run_img

# run container with docker volume to mount models and config folder
make docker_cpu_debian_lam_audio2expression_avatar_container_run

run websocket signaling service + webrtc + websocket voice avatar agent web ui

cd ui/webrtc_websocket/lam_audio2expression_avatar_ts_v2 && npm install && npm run dev

More details: #164 and #206 | online lam_audio2expression avatar: https://avatar-2lm.pages.dev/

---

Run chat bots

• 📝 Run chat bots with colab notebook

License

achatbot is released under the BSD 3 license. (Additional code in this distribution is covered by the MIT and Apache Open Source licenses.) However you may have other legal obligations that govern your use of content, such as the terms of service for third-party models.