https://github.com/libre-webui/libre-webui

Privacy-first web interface for local AI models. Clean, minimal UI for Ollama with extensible plugin system. No data leaves your device.
https://github.com/libre-webui/libre-webui

ai ai-interface chat-interface enterprise-services foss libre llm local-ai minimal-ui nodejs ollama open-source plugin-system privacy-first react self-hosted typescript ui webui

Last synced: 7 days ago
JSON representation

Privacy-first web interface for local AI models. Clean, minimal UI for Ollama with extensible plugin system. No data leaves your device.

Host: GitHub
URL: https://github.com/libre-webui/libre-webui
Owner: libre-webui
License: apache-2.0
Created: 2025-05-19T15:16:10.000Z (9 months ago)
Default Branch: main
Last Pushed: 2026-01-31T23:37:50.000Z (10 days ago)
Last Synced: 2026-02-01T08:23:15.688Z (9 days ago)
Topics: ai, ai-interface, chat-interface, enterprise-services, foss, libre, llm, local-ai, minimal-ui, nodejs, ollama, open-source, plugin-system, privacy-first, react, self-hosted, typescript, ui, webui
Language: TypeScript
Homepage: https://librewebui.org
Size: 11.6 MB
Stars: 22
Watchers: 2
Forks: 7
Open Issues: 0
Metadata Files:
- Readme: README.md
- Changelog: CHANGELOG.md
- Funding: .github/FUNDING.yml
- License: LICENSE

Awesome Lists containing this project

README

          


# Libre WebUI

### Privacy-First AI Chat Interface



  Self-hosted • Open Source • Extensible


  _{Enterprise support by Kroonen AI}





  





  

  

  

  





  

  

  



[Website](https://librewebui.org) • [Documentation](https://docs.librewebui.org) • [GitLab](https://git.kroonen.ai/libre-webui/libre-webui) • [𝕏](https://x.com/librewebui) • [Sponsor](https://github.com/sponsors/libre-webui) • [Get Started](#quick-start)



---

## Why Libre WebUI?

A simple, self-hosted interface for AI chat. Run it locally with Ollama, connect to OpenAI, Anthropic, HuggingFace, or 10+ providers—all from one UI.

- **Your data stays yours** — Zero telemetry, fully self-hosted

- **Extensible plugin system** — Ollama, OpenAI, Anthropic, and any OpenAI-compatible API

- **Simple & focused** — Keyboard shortcuts, dark mode, responsive design

---

## Features

### Core Experience

- Real-time streaming chat

- Dark/light themes

- VS Code-style keyboard shortcuts

- Mobile-responsive design

- **Native Desktop App** — macOS (Windows & Linux coming soon)

### AI Providers

- **Local**: Ollama (full integration)

- **Cloud**: OpenAI, Anthropic, Google, Groq, Mistral, OpenRouter, HuggingFace, and more

- **HuggingFace Hub** — 1M+ models for chat, TTS, image gen, embeddings, STT

- **Image Generation** — ComfyUI with Flux models

- **Plugin System** — Add any OpenAI-compatible API via JSON config

- **Plugin Variables** — Per-plugin configurable settings (temperature, endpoint, etc.)

### Advanced Capabilities

- **Document Chat (RAG)** — Upload PDFs, chat with your docs

- **Custom Personas** — AI personalities with memory

- **Interactive Artifacts** — Live HTML, SVG, code preview

- **Text-to-Speech** — Multiple voices and providers

- **SSO Authentication** — GitHub, Hugging Face OAuth

### Security

- AES-256-GCM encryption

- Role-based access control

- Enterprise compliance ready

---

## Quick Start

**Requirements:** [Ollama](https://ollama.ai) (for local AI) or API keys for cloud providers

### One Command Install

```bash

npx libre-webui

```

That's it. Opens at `http://localhost:8080`

### Homebrew (macOS)

```bash

# CLI version (includes backend server)

brew tap libre-webui/tap

brew install libre-webui

libre-webui

# Or desktop app

brew install --cask libre-webui

```

Run as a background service:

```bash

brew services start libre-webui

```

### Docker

| Setup                                     | Command                                                      |

| ----------------------------------------- | ------------------------------------------------------------ |

| Bundled Ollama (CPU)                      | `docker-compose up -d`                                       |

| Bundled Ollama (NVIDIA GPU)               | `docker-compose -f docker-compose.gpu.yml up -d`             |

| External Ollama (already running on host) | `docker-compose -f docker-compose.external-ollama.yml up -d` |

Access at `http://localhost:8080`

Development builds (unstable)

> **Warning:** Development builds are automatically generated from the `dev` branch and may contain experimental features, breaking changes, or bugs. Use at your own risk and do not use in production environments.

| Setup                             | Command                                                          |

| --------------------------------- | ---------------------------------------------------------------- |

| Dev + Bundled Ollama (CPU)        | `docker-compose -f docker-compose.dev.yml up -d`                 |

| Dev + Bundled Ollama (NVIDIA GPU) | `docker-compose -f docker-compose.dev.gpu.yml up -d`             |

| Dev + External Ollama             | `docker-compose -f docker-compose.dev.external-ollama.yml up -d` |

Development builds use separate data volumes (`libre_webui_dev_data`) to prevent conflicts with stable installations.

To pull the latest dev image manually:

```bash

docker pull librewebui/libre-webui:dev

```

### Kubernetes (Helm)

```bash

helm install libre-webui oci://ghcr.io/libre-webui/charts/libre-webui

```

Helm configuration options

```bash

# With external Ollama

helm install libre-webui oci://ghcr.io/libre-webui/charts/libre-webui \

  --set ollama.bundled.enabled=false \

  --set ollama.external.enabled=true \

  --set ollama.external.url=http://my-ollama:11434

# With NVIDIA GPU support

helm install libre-webui oci://ghcr.io/libre-webui/charts/libre-webui \

  --set ollama.bundled.gpu.enabled=true

# With Ingress

helm install libre-webui oci://ghcr.io/libre-webui/charts/libre-webui \

  --set ingress.enabled=true \

  --set ingress.hosts[0].host=chat.example.com

```

See [helm/libre-webui/values.yaml](helm/libre-webui/values.yaml) for all configuration options.

### Development Setup

```bash

# 1. Clone the repo

git clone https://github.com/libre-webui/libre-webui

cd libre-webui

# 2. Configure environment

cp backend/.env.example backend/.env

# 3. Install and run

npm install && npm run dev

```

### Configuration

Edit `backend/.env` to add your API keys:

```env

# Local AI (Ollama)

OLLAMA_BASE_URL=http://localhost:11434

# Cloud AI Providers (add the ones you need)

OPENAI_API_KEY=sk-...

ANTHROPIC_API_KEY=sk-ant-...

HUGGINGFACE_API_KEY=hf_...

```

---

## Plugin System

Plugins let you connect any OpenAI-compatible (or custom) API to Libre WebUI using a simple JSON file. Built-in plugins are included for OpenAI, Anthropic, Google Gemini, Groq, Mistral, OpenRouter, HuggingFace, and more.

### How It Works

A plugin is a JSON config file stored in the `plugins/` directory. Each plugin defines:

- **Endpoint** — The API URL to send requests to

- **Auth** — How to authenticate (header name, prefix, env variable fallback)

- **Model map** — Which models the plugin supports

- **Capabilities** — Optional multi-capability support (chat, TTS, image generation, embeddings, STT)

- **Variables** — Per-plugin configurable settings users can adjust from the UI

### Example Plugin

```json

{

  "id": "my-provider",

  "name": "My Provider",

  "type": "completion",

  "endpoint": "https://api.example.com/v1/chat/completions",

  "auth": {

    "header": "Authorization",

    "prefix": "Bearer ",

    "key_env": "MY_PROVIDER_API_KEY"

  },

  "model_map": ["model-a", "model-b"],

  "variables": [

    {

      "name": "temperature",

      "type": "number",

      "label": "Temperature",

      "description": "Controls randomness.",

      "default": 0.7,

      "min": 0,

      "max": 2

    }

  ]

}

```

### Plugin Management

Plugins are managed from **Settings > Plugins** in the UI where you can:

- **Upload** a JSON plugin file or paste JSON directly

- **Browse HuggingFace** to discover and import models

- **Activate/deactivate** plugins

- **Set API keys** per plugin (encrypted at rest with AES-256-GCM, with env variable fallback)

- **Configure variables** — override endpoint, temperature, max tokens, and other settings per plugin

- **Export** plugins as JSON

### Plugin Variables

Variables are typed settings defined in the plugin JSON that users can configure from the UI. Supported types: `string`, `number`, `boolean`, `select`. Sensitive variables (like API keys) are encrypted in the database. Variable values are stored per-user.

### Multi-Capability Plugins

A single plugin can support multiple capabilities. For example, OpenAI's plugin handles both chat completions and TTS:

```json

{

  "id": "openai",

  "type": "completion",

  "endpoint": "https://api.openai.com/v1/chat/completions",

  "capabilities": {

    "tts": {

      "endpoint": "https://api.openai.com/v1/audio/speech",

      "model_map": ["tts-1", "tts-1-hd"],

      "config": {

        "voices": ["alloy", "echo", "nova", "shimmer"],

        "default_voice": "alloy"

      }

    }

  }

}

```

### Creating Custom Plugins

To add a new provider, create a JSON file in the `plugins/` directory following the schema above, or upload one through the UI. Any OpenAI-compatible API works out of the box — just set the correct endpoint and auth config.

For the full plugin reference (variables, credentials, security, API endpoints), see the [Plugin Architecture docs](./docs/08-PLUGIN_ARCHITECTURE.md).

---

### Desktop App (In Development)

> **Note:** The desktop app is currently in active development. The macOS build is pending Apple notarization, which may cause security warnings or installation issues on some systems. We're working to resolve this. Feedback and bug reports are welcome!

Download the native desktop app from [GitHub Releases](https://github.com/libre-webui/libre-webui/releases):

| Platform              | Status                  |

| --------------------- | ----------------------- |

| macOS (Apple Silicon) | Beta (`.dmg` or `.zip`) |

| Windows               | Coming soon             |

| Linux                 | Coming soon             |

---

> [!NOTE]

>

> ## Enterprise Services

>

> **Need a custom deployment?** [Kroonen AI](https://kroonen.ai) provides professional services for Libre WebUI deployments.

>

> | Service                       | Use Case                              |

> | ----------------------------- | ------------------------------------- |

> | On-premise & cloud deployment | HIPAA, SOC 2, air-gapped environments |

> | SSO integration               | Okta, Azure AD, SAML, LDAP            |

> | Custom development            | Integrations, white-labeling, plugins |

> | SLA-backed support            | Priority response, dedicated channel  |

>

> **Contact:** enterprise@kroonen.ai | **[Learn more →](https://kroonen.ai/services)**

> [!TIP]

>

> ## Support Development

>

> Libre WebUI is built and maintained independently. Your support keeps it free and open source.

>

> [![Sponsor](https://img.shields.io/badge/Sponsor-❤️-red?style=for-the-badge&logo=github)](https://github.com/sponsors/libre-webui)

>

> **[Become a Sponsor](https://github.com/sponsors/libre-webui)** — Help fund active development

---

## Community

- [Ethical Charter](./CHARTER.md) — Our commitment to privacy, freedom & transparency

- [Contributing](https://github.com/libre-webui/libre-webui/contribute) — Help improve Libre WebUI

- [𝕏 @librewebui](https://x.com/librewebui) — Follow for updates

- [Mastodon](https://fosstodon.org/@librewebui) — Fediverse updates

- [GitLab](https://git.kroonen.ai/libre-webui/libre-webui) — Self-hosted mirror

- [GitHub Issues](https://github.com/libre-webui/libre-webui/issues) — Bug reports & feature requests

- [Documentation](https://docs.librewebui.org) — Guides & API reference

---



**Apache 2.0 License** • Copyright © 2025–present Libre WebUI™

Built & maintained by [Kroonen AI](https://kroonen.ai) • [Enterprise Support](https://kroonen.ai/services)

ecosyste.ms

Data

Tools

Indexes

Applications

Experiments

Awesome

https://github.com/libre-webui/libre-webui

Awesome Lists containing this project

README