[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-72290":3},{"id":4,"name":5,"fullName":6,"owner":7,"repo":5,"description":8,"homepage":8,"htmlUrl":8,"language":9,"languages":8,"totalLinesOfCode":8,"stars":10,"forks":11,"watchers":12,"openIssues":13,"contributorsCount":14,"subscribersCount":14,"size":14,"stars1d":15,"stars7d":16,"stars30d":17,"stars90d":14,"forks30d":14,"starsTrendScore":18,"compositeScore":19,"rankGlobal":8,"rankLanguage":8,"license":20,"archived":21,"fork":21,"defaultBranch":22,"hasWiki":23,"hasPages":23,"topics":24,"createdAt":8,"pushedAt":8,"updatedAt":25,"readmeContent":26,"aiSummary":27,"trendingCount":14,"starSnapshotCount":14,"syncStatus":28,"lastSyncTime":29,"discoverSource":30},72290,"OpenAvatarChat","HumanAIGC-Engineering\u002FOpenAvatarChat","HumanAIGC-Engineering",null,"Python",3534,582,37,148,0,11,36,117,33,30.3,"Apache License 2.0",false,"main",true,[],"2026-06-12 02:03:01","\u003Ch1 style='text-align: center; margin-bottom: 1rem'> Open Avatar Chat \u003C\u002Fh1>\n\n\u003Cp align=\"center\">\n\u003Cstrong>中文 | \u003Ca href=\"readme_en.md\">English\u003C\u002Fa>\u003C\u002Fstrong>\n\u003C\u002Fp>\n\n\u003Cp align=\"center\">\n\u003Cstrong>模块化的交互数字人对话实现。\u003C\u002Fstrong>\n\u003C\u002Fp>\n\n\u003Cp align=\"center\" style=\"display: flex; flex-direction: row; justify-content: center\">\n 🤗 \u003Ca href=\"https:\u002F\u002Fhuggingface.co\u002Fspaces\u002FHumanAIGC-Engineering-Team\u002Fopen-avatar-chat\">Demo\u003C\u002Fa>&nbsp&nbsp|&nbsp&nbsp\u003Cimg alt=\"Static Badge\" style=\"height: 10px;\" src=\".\u002Fassets\u002Fimages\u002Fmodelscope_logo.png\"> \u003Ca href=\"https:\u002F\u002Fwww.modelscope.cn\u002Fstudios\u002FHumanAIGC-Engineering\u002Fopen-avatar-chat\">Demo\u003C\u002Fa>&nbsp&nbsp|&nbsp&nbsp💬 \u003Ca href=\"https:\u002F\u002Fgithub.com\u002FHumanAIGC-Engineering\u002FOpenAvatarChat\u002Fblob\u002Fmain\u002Fassets\u002Fimages\u002Fcommunity_wechat.png\">WeChat (微信)\u003C\u002Fa>&nbsp&nbsp|&nbsp&nbsp📖 \u003Ca href=\"https:\u002F\u002Fhumanaigc-engineering.github.io\u002FOpenAvatarChat\u002F\">文档\u003C\u002Fa>\n\u003C\u002Fp>\n\n## 💡 核心亮点\n\n- **多模态交互支持**：支持文本、语音、视频等多种交互方式，提供自然流畅的人机对话体验\n- **模块化架构设计**：采用高度模块化设计，可灵活替换 ASR、LLM、TTS、Avatar 等核心组件\n- **多样数字人形象**：支持 LiteAvatar、LAM、MuseTalk、FlashHead 等多种数字人技术\n- **低延迟优化**：通过 VAD 检测、语音缓冲、帧率控制等机制优化，平均响应时间仅 2.2 秒\n\n## 📢 最新动态\n\n- [2026.04] ⭐️⭐️⭐️ 版本 0.6.0发布:\n  - 架构重构，前后端分离，前端仓库 [OpenAvatarChat-WebUI](https:\u002F\u002Fgithub.com\u002FHumanAIGC-Engineering\u002FOpenAvatarChat-WebUI)\n  - 所有数字人均支持手动打断和双工打断模式\n  - 优化安装部署和模型下载流程，统一依赖管理和模型下载脚本\n  - 接入 [SoulX-FlashHead](https:\u002F\u002Fgithub.com\u002FSoul-AILab\u002FSoulX-FlashHead) 数字人，基于扩散模型的实时流式说话头生成\n- [2025.08.19] ⭐️⭐️⭐️ 版本 0.5.1发布:\n  - LiteAvatar支持单机多session\n  - 增加对 Qwen-Omni多模态模型的支持\n\n> 📋 [完整更新日志](https:\u002F\u002Fhumanaigc-engineering.github.io\u002FOpenAvatarChat\u002Freleases\u002Frelease-notes)\n\n## Demo\n\n### 在线体验\n\n我们部署在\n\u003Ca href=\"https:\u002F\u002Fwww.modelscope.cn\u002Fstudios\u002FHumanAIGC-Engineering\u002Fopen-avatar-chat\" target=\"_blank\">ModelScope\u003C\u002Fa>\n和\n\u003Ca href=\"https:\u002F\u002Fhuggingface.co\u002Fspaces\u002FHumanAIGC-Engineering-Team\u002Fopen-avatar-chat\" target=\"_blank\">HuggingFace\u003C\u002Fa>\n上均部署了体验服务，欢迎体验。\n\n### 视频\n\u003Ctable>\n  \u003Ctr>\n    \u003Ctd align=\"center\">\n      \u003Ch3>LiteAvatar\u003C\u002Fh3>\n      \u003Cvideo controls src=\"https:\u002F\u002Fgithub.com\u002Fuser-attachments\u002Fassets\u002Fe2861200-84b0-4c7a-93f0-f46268a0878b\">\u003C\u002Fvideo>\n    \u003C\u002Ftd>\n    \u003Ctd align=\"center\">\n      \u003Ch3>LAM\u003C\u002Fh3>\n      \u003Cvideo controls src=\"https:\u002F\u002Fgithub.com\u002Fuser-attachments\u002Fassets\u002Fa72a8c33-39dd-4656-a4a9-b76c5487c711\">\u003C\u002Fvideo>\n    \u003C\u002Ftd>\n  \u003C\u002Ftr>\n\u003C\u002Ftable>\n\n## 组件依赖\n\n| 类型       | 开源项目                                |Github地址|模型地址|\n|----------|-------------------------------------|---|---|\n| RTC      | HumanAIGC-Engineering\u002Fgradio-webrtc |[\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fgithub-white?logo=github&logoColor=black\"\u002F>](https:\u002F\u002Fgithub.com\u002FHumanAIGC-Engineering\u002Fgradio-webrtc)||\n| WebUI      | HumanAIGC-Engineering\u002FOpenAvatarChat-WebUI |[\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fgithub-white?logo=github&logoColor=black\"\u002F>](https:\u002F\u002Fgithub.com\u002FHumanAIGC-Engineering\u002FOpenAvatarChat-WebUI)||\n| VAD      | snakers4\u002Fsilero-vad                 |[\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fgithub-white?logo=github&logoColor=black\"\u002F>](https:\u002F\u002Fgithub.com\u002Fsnakers4\u002Fsilero-vad)||\n| Avatar   | HumanAIGC\u002Flite-avatar               |[\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fgithub-white?logo=github&logoColor=black\"\u002F>](https:\u002F\u002Fgithub.com\u002FHumanAIGC\u002Flite-avatar)||\n| TTS      | FunAudioLLM\u002FCosyVoice               |[\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fgithub-white?logo=github&logoColor=black\"\u002F>](https:\u002F\u002Fgithub.com\u002FFunAudioLLM\u002FCosyVoice)||\n|Avatar|aigc3d\u002FLAM_Audio2Expression|[\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fgithub-white?logo=github&logoColor=black\"\u002F>](https:\u002F\u002Fgithub.com\u002Faigc3d\u002FLAM_Audio2Expression)|[🤗](https:\u002F\u002Fhuggingface.co\u002F3DAIGC\u002FLAM_audio2exp)|\n||facebook\u002Fwav2vec2-base-960h||[🤗](https:\u002F\u002Fhuggingface.co\u002Ffacebook\u002Fwav2vec2-base-960h)  [\u003Cimg src=\".\u002Fassets\u002Fimages\u002Fmodelscope_logo.png\" width=\"20px\">\u003C\u002Fimg>](https:\u002F\u002Fmodelscope.cn\u002Fmodels\u002FAI-ModelScope\u002Fwav2vec2-base-960h)|\n|Avatar|TMElyralab\u002FMuseTalk|[\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fgithub-white?logo=github&logoColor=black\"\u002F>](https:\u002F\u002Fgithub.com\u002FTMElyralab\u002FMuseTalk)||\n|Avatar|Soul-AILab\u002FSoulX-FlashHead|[\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fgithub-white?logo=github&logoColor=black\"\u002F>](https:\u002F\u002Fgithub.com\u002FSoul-AILab\u002FSoulX-FlashHead)|[🤗](https:\u002F\u002Fhuggingface.co\u002FSoul-AILab\u002FSoulX-FlashHead-1_3B)|\n||||||\n\n## 🚀 快速开始\n\n```bash\n# 克隆项目\ngit clone https:\u002F\u002Fgithub.com\u002FHumanAIGC-Engineering\u002FOpenAvatarChat.git\ncd OpenAvatarChat\ngit submodule update --init --recursive --depth 1\n\n# 安装依赖（以 LiteAvatar + 百炼 API 为例）\nuv run install.py --config config\u002Fchat_with_openai_compatible_bailian_cosyvoice.yaml\n\n# 下载模型\nuv run scripts\u002Fdownload_models.py --handler liteavatar\n\n# 启动\nuv run src\u002Fdemo.py --config config\u002Fchat_with_openai_compatible_bailian_cosyvoice.yaml\n```\n\n> 📖 详细步骤请参阅[快速开始文档](https:\u002F\u002Fhumanaigc-engineering.github.io\u002FOpenAvatarChat\u002Fgetting-started\u002F)\n\n## 预置模式\n\n| CONFIG名称 | ASR | LLM | TTS | AVATAR |\n|----------------------------------------------------|-----|:---------:|:---------:|------------|\n| chat_with_lam.yaml | SenseVoice | API | API | LAM |\n| chat_with_qwen_omni.yaml | Qwen-Omni | Qwen-Omni | Qwen-Omni | lite-avatar |\n| chat_with_openai_compatible_bailian_cosyvoice.yaml | SenseVoice | API | API | lite-avatar |\n| chat_with_openai_compatible_bailian_cosyvoice_flashhead.yaml | SenseVoice | API | API | FlashHead |\n| chat_with_openai_compatible_bailian_cosyvoice_flashhead_duplex.yaml | SenseVoice | API | API | FlashHead (双工) |\n| chat_with_openai_compatible_bailian_cosyvoice_flashhead_duplex_agent.yaml | SenseVoice | **Agent** | API | FlashHead (双工+Agent) Beta |\n\n> 📖 [查看全部预置模式](https:\u002F\u002Fhumanaigc-engineering.github.io\u002FOpenAvatarChat\u002Freference\u002Fpreset-modes)\n\n## 🧪 Beta 功能\n\n### Chat Agent 模式（OpenClaw 集成）\n\n> [!WARNING]\n> 此功能目前处于 **Beta** 阶段，API 和配置格式可能随时变化。\n\nChat Agent 模式使用多轮工具调用 Agent 替代传统 LLM Handler，为数字人提供：\n\n- **工具调用**：多轮调用工具（获取时间、系统信息等）\n- **人格与长期记忆**：通过 OpenClaw 的 Agent Profile 赋予数字人持久人格\n- **对话上下文压缩**：自动压缩过长的对话历史\n- **后台任务协作**：通过 OpenClaw 在后台执行复杂任务\n- **视觉感知**：结合 PerceptionAgent 处理摄像头输入\n\n> 📖 [查看完整 Chat Agent 文档](https:\u002F\u002Fhumanaigc-engineering.github.io\u002FOpenAvatarChat\u002Fbeta\u002Fchat-agent)\n\n## 社区\n\n* 微信群\n\n\u003Cimg alt=\"community_wechat.png\" height=\"200\" src=\"https:\u002F\u002Fgithub.com\u002FHumanAIGC-Engineering\u002FOpenAvatarChat\u002Fblob\u002Fmain\u002Fassets\u002Fimages\u002Fcommunity_wechat.png\" width=\"200\"\u002F>\n\n* 官方视频教程：[Bilibili](https:\u002F\u002Fwww.bilibili.com\u002Fvideo\u002FBV1sv8QzLEC2)\n* 🚨 [常见问题](https:\u002F\u002Fhumanaigc-engineering.github.io\u002FOpenAvatarChat\u002Fcommunity\u002Ffaq)\n\n## 社区贡献-感谢\n\n- 感谢社区热心同学\"十字鱼\"在B站上发布的一键安装包视频 [一键包](https:\u002F\u002Fwww.bilibili.com\u002Fvideo\u002FBV1V1oLYmEu3\u002F?vd_source=29463f5b63a3510553325ba70f325293)\n- 感谢社区热心同学\"W&H\"提供的夸克一键包[windows版本:提取码a79V](https:\u002F\u002Fpan.quark.cn\u002Fs\u002F237177126010) 和 [linux 版本:提取码：E8Kq](https:\u002F\u002Fpan.quark.cn\u002Fs\u002Fb7fcdc157586)\n- 感谢社区热心同学\"W&H\"提供的源码zip[夸克网盘:提取码 9iNy](https:\u002F\u002Fpan.quark.cn\u002Fs\u002F9e6156cafacd) 和 [百度云盘:提取码：xrxr](https:\u002F\u002Fpan.baidu.com\u002Fs\u002F16-0OBtSD5cBz2gJDJORW7w)\n\n## Star历史\n\n![](https:\u002F\u002Fapi.star-history.com\u002Fsvg?repos=HumanAIGC-Engineering\u002FOpenAvatarChat&type=Date)\n\n## 引用\n\n如果您在您的研究\u002F项目中感到 OpenAvatarChat 为您提供了帮助，期待您能给一个 Star⭐和引用✏️\n\n```\n@software{avatarchat2025,\n  author = {Gang Cheng, Tao Chen, Feng Wang, Binchao Huang, Hui Xu, Guanqiao He, Yi Lu, Shengyin Tan},\n  title = {OpenAvatarChat},\n  year = {2025},\n  publisher = {GitHub},\n  url = {https:\u002F\u002Fgithub.com\u002FHumanAIGC-Engineering\u002FOpenAvatarChat}\n}\n```\n","OpenAvatarChat 是一个模块化的交互数字人对话实现项目。它支持多模态交互，包括文本、语音和视频，提供自然流畅的人机对话体验。该项目采用高度模块化设计，允许用户灵活替换自动语音识别（ASR）、大语言模型（LLM）、文本转语音（TTS）以及数字人形象等核心组件。此外，OpenAvatarChat 支持多种数字人技术，如LiteAvatar、LAM、MuseTalk和FlashHead，并通过VAD检测、语音缓冲及帧率控制等机制优化了响应时间，平均响应时间仅为2.2秒。此项目适用于需要高质量人机交互的应用场景，如虚拟客服、在线教育、娱乐互动等领域。",2,"2026-06-11 03:41:12","high_star"]