开源 & 自托管
你的 AI，你的规则
v0.11.0 • 最新动态
发布说明 查看全部 →
✨ New Features
- Added optional Cloudflare Turnstile protection for public account creation, configured from backend environment variables and rendered only when enabled.
- Added Tailscale-friendly development hosting/origin support so Libre WebUI can be reached from another device during local development.
- Added Playwright e2e coverage for demo login, one-user mode, artifact detection/rendering, artifact resize behavior, cloud-model regression coverage, mobile sidebar behavior, and async locale loading.
🔧 Improvements
- Set blue as the default accent color and polished the remaining auth, artifact demo, badge, icon, logo, and model-flow surfaces to follow the current design system.
- Improved artifact reliability for multi-file HTML, filename-qualified bundles, standalone full HTML documents, local file reference cleanup, and themed empty/fallback preview states.
- Made artifact pane resizing release pointer state correctly so the pane no longer feels stuck to the mouse after resize.
- Made code-block streaming smoother with requestAnimationFrame-coalesced updates, a lightweight streaming code renderer, lazy rich markdown rendering, and split markdown/math/highlighting chunks.
- Reworked title generation so it honors the existing toggle, resolves the current running model correctly, supports plugin-backed title generation, falls back to Ollama, and sanitizes generated titles.
- Split major backend chat orchestration into focused services for WebSocket handling, shared chat context, assistant completion persistence, plugin streaming, Ollama streaming, title generation, and request preparation.
- Split plugin capabilities into smaller adapters/services for chat, streaming, embeddings, image generation, TTS, validation, uploads, variables, and provider response conversion.
- Slimmed large frontend surfaces by extracting Settings tabs, Model Manager sections, Model Selector tabs, Persona Form tabs, Sidebar sections, API domain clients, and chat store helpers.
- Added gated frontend/backend loggers and routed noisy development diagnostics through explicit debug controls.
- Updated release tooling with a security preflight and kept the production build free of the previous large chunk warnings.
🐛 Bug Fixes
- Fixed toggled session title generation so titles are generated only when the option is enabled.
- Fixed demo mode persona avatars so bundled/local demo images render correctly.
- Fixed chat routing and settings state edge cases after the shared generation refactor.
- Fixed dark-mode artifact demo surfaces so generated previews no longer fall back to the old default blue styling.
- Added regression coverage around cloud model suffix handling without re-documenting the 0.10.0 cloud model feature itself.
🔒 Security & Dependencies
- Remediated npm security advisories and refreshed dependency locks after 0.10.0, including updates around
multer,undici,ws,ip-address,brace-expansion,tmp, andqs. - Added
npm auditto the release preflight path and verified the package lock reports zero vulnerabilities after the release update. - Updated GitHub Actions dependencies, including
actions/checkout, and hardened the release workflow checks before publishing.
📚 Documentation
- Refreshed the README to describe the current Libre WebUI product, model/provider story, install paths, and project positioning.
- Audited and rewrote the project docs so the model examples, setup guidance, plugin docs, auth docs, artifacts docs, environment variables, and deployment guides match the current application.
- Updated DESIGN.md with the current blue-accent direction and UI guidance without repeating the 0.10.0 custom-accent release notes.
⚠️ Breaking Changes
- No known user-facing breaking changes. The plugin and chat internals were heavily refactored, but existing plugin configuration and normal app workflows remain compatible.
隐私优先的 AI 聊天界面。使用 Ollama 本地运行，或连接 OpenAI、Anthropic 等 9+ 服务商。 零遥测。零追踪。
npx libre-webui 需要 Node.js 18+ 和 Ollama （用于本地 AI）
零遥测
自托管
Apache 2.0
插件系统
你需要的一切
尊重隐私的完整 AI 聊天解决方案
交互式工件
交互式工件
直接在聊天中渲染 HTML、SVG 和 React 组件。支持全屏模式实时预览。
本地 & 云端 AI 文档对话 (RAG) 插件变量
本地 & 云端 AI
使用 Ollama 本地运行模型，或连接 OpenAI、Anthropic、Groq、Gemini、Mistral 等。选择权在你。
文档对话 (RAG)
上传 PDF、文档和文本文件。通过语义搜索和向量嵌入询问文档相关问题。
交互式工件
直接在聊天中渲染 HTML、SVG 和 React 组件。支持全屏模式实时预览。
AES-256 加密
为所有数据提供企业级加密。聊天记录、文档和设置均加密存储。
自定义角色
创建具有独特行为和系统提示的 AI 个性。支持 JSON 导入/导出角色。
文字转语音
使用多种语音选项收听 AI 回复。支持浏览器 TTS 和 ElevenLabs 集成。
图像生成
使用 ComfyUI 和 Flux 模型生成图像。多种尺寸、质量预设，无缝集成到聊天中。
插件变量
每个插件可配置的设置，如温度、端点和令牌。敏感值会加密存储。
键盘快捷键
为高级用户提供 VS Code 风格快捷键。通过键盘导航、切换设置、控制一切。
多用户支持
支持 SSO 的基于角色的访问控制。内置 GitHub 和 Hugging Face OAuth。
连接任何服务商
一个界面，无限可能
Ollama
本地模型
OpenAI
GPT-4o, o1, o3
Anthropic
Claude 4, Opus
Groq
Llama, Mixtral
Google
Gemini Pro
Mistral
Mistral Large
OpenRouter
400+ 模型
HuggingFace
1M+ models
+ Custom
任何 OpenAI 兼容 API
几秒钟即可开始
选择你喜欢的安装方式
创建自定义插件
通过简单的 JSON 文件连接任何 OpenAI 兼容 LLM
可用插件
Libre WebUI 仓库的官方插件。点击查看或下载。
从 GitHub 加载插件...
custom-model.json
{
"id": "custom-model",
"name": "Custom Model",
"type": "completion",
"endpoint": "http://localhost:8000/v1/chat/completions",
"auth": {
"header": "Authorization",
"prefix": "Bearer ",
"key_env": "CUSTOM_MODEL_API_KEY"
},
"model_map": [
"my-fine-tuned-llama"
],
"variables": [
{
"name": "temperature",
"type": "number",
"label": "Temperature",
"default": 0.7,
"min": 0, "max": 2
}
]
}创建你自己的插件
1
启动 LLM 服务器
运行任何 OpenAI 兼容服务器：llama.cpp、vLLM、Ollama 或自定义 FastAPI 服务器。
2
创建插件 JSON
在简单的 JSON 文件中定义端点、认证和可用模型。
3
上传到 Libre WebUI
进入设置 > 服务商，上传插件并输入 API 密钥。
4
开始聊天
自定义模型将出现在模型选择器中。完全隐私，完全掌控。
插件字段参考
id 唯一标识符（小写，可用连字符） name UI 中显示的名称 type 聊天用 "completion"，语音合成用 "tts" endpoint API URL（例如：/v1/chat/completions） auth.header 认证头名称（Authorization、x-api-key） auth.prefix 密钥前缀（"Bearer " 或空） auth.key_env API 密钥的环境变量 model_map 可用模型标识符数组 variables 可配置的设置（温度、端点等） 准备好掌控你的 AI 了吗？
加入数千名重视隐私和掌控的用户。