Open WebUI
Open WebUI is a home for AI, a self-hosted AI platform that's extensible, feature-rich, user-friendly, and built to run entirely offline. With support for Ollama and OpenAI-compatible APIs, it gives you a powerful, provider-agnostic interface for both local and cloud-based models.
Quick Start
- Docker
- pip
- uv
- Desktop
docker run -d -p 3000:8080 --add-host=host.docker.internal:host-gateway -v open-webui:/app/backend/data -e WEBUI_SECRET_KEY=your-secret-key --name open-webui --restart always ghcr.io/open-webui/open-webui:mainThen open http://localhost:3000.
Powered by Open Terminal and some of our amazing community built plugins.
For GPU support, Docker Compose, and more → Quick Start and Docker images and tags
:main above is the batteries-included one: it carries the embedding, speech and reranking models, document extraction and every database client, so retrieval, uploads and voice work the moment it starts.
:main-slim is a fraction of that size and leaves the machine learning stack out. Chatting with a hosted model is identical, and the features that used local models run on services you point it at instead. Pick it when you want a small install, few dependencies, or Open WebUI purely as a front end. See What slim leaves out.
pip install open-webui
open-webui serveThen open http://localhost:8080.
curl -LsSf https://astral.sh/uv/install.sh | sh
DATA_DIR=~/.open-webui uvx --python 3.11 open-webui@latest serveThen open http://localhost:8080.
Download the desktop app from github.com/open-webui/desktop. It runs Open WebUI natively on your system without Docker or manual setup.
For production deployments, install via Docker or Python.
Installed Open WebUI but not sure where to start? The Essentials for Open WebUI guide covers what every new user needs to know: plugins, tool calling, task models, context management, RAG, web search and Open Terminal.
Getting Started
- Quick Start: Docker, Python, Kubernetes and desktop app install options
- Connect a Provider: Ollama, OpenAI, Anthropic, vLLM, and more
- Essentials for Open WebUI: Start here after your first install. Plugins, tool calling, task models, context management, RAG, web search, Open Terminal.
- Connect an Agent: Open WebUI Computer, Hermes Agent, OpenClaw, and other autonomous AI agents
- Updating: Keep your instance current
- Development Branch: Help test the latest changes before stable release
- Advanced Topics: Scaling, logging, and advanced configuration
Explore
- Features: Discover what Open WebUI can do
- Tutorials: Step-by-step guides
- FAQ: Common questions answered
- Troubleshooting: Fix common issues
- Reference: Environment variables and API details
Agentic Execution with Open Terminal
Open Terminal gives an agent a real terminal and filesystem, so it can carry out multi-step work instead of only describing it: analyze data, run scripts, fix its own errors, and hand you the files it produced, all inside the chat. A file browser beside the conversation shows everything it creates.
For teams, Terminals is the enterprise layer on top, giving every user an isolated environment with resource limits and automatic lifecycle management.
Open WebUI Computer
Open WebUI Computer is the cutting-edge agent harness and runtime for your real machine. You install one app, add AI (an API key, a local model, or a coding-agent subscription you already have like Claude Code or Codex), and it works on the actual files, branches, services, and tools already on that computer, with approvals and plan mode. The file the agent edits is the file on your disk; the terminal it uses is the terminal on your machine.
Run Computer with uv:
uvx cptr@latest runThen create an account and open a folder.
Because it's your whole machine, you get the rest for free: files, editor, terminal, git, previews, and agent chats in any browser, from your desk or your phone; an assistant you can message from Telegram or WhatsApp; scheduled agents that report back to you. Everything runs locally on hardware you own, nothing hosted elsewhere.
Computer and Open WebUI are one ecosystem moving at two speeds, on purpose. Open WebUI is the stable platform: enterprise-ready, multi-user, what teams and organizations deploy and rely on. Computer moves faster: new real-machine agent capabilities are built and shipped there first, hardened in daily use, and graduated into Open WebUI where they fit. If you want the bleeding edge of what we're building, it lands in Computer first.
- New to it? What is Computer? · Quickstart · Use cases
- Already run Open WebUI? Use a Computer workspace as a model in Open WebUI
Also by Open WebUI
- open-terminal: Turn Open WebUI into a full agent harness with a real action runtime: sandboxed files, shell, code execution, previews, and artifacts.
- oikb: Keep your Open WebUI knowledge bases in sync. Watches local folders, GitHub repos, S3 buckets, Confluence, and 40+ other sources.
- mcpo: MCP-to-OpenAPI proxy. Use any MCP tool server with Open WebUI (or any OpenAPI client) without custom glue code.
Enterprise
Built for teams and organizations, Open WebUI combines full control over your data and infrastructure with a stateless architecture that scales from a pilot to an organization-wide deployment.
Need custom branding, SLA support, or Long-Term Support (LTS)? → Learn about Enterprise plans
Get Involved
- Contributing: Help build Open WebUI
- Development Setup: Run the project locally from source
- Discord: Join the community
- GitHub: Report issues, submit PRs
- Careers: Join our team
Sponsors
Acknowledgements
For LLMs and coding agents
Machine-readable entry points to this documentation:
- llms.txt: curated index of every docs page with short descriptions.
- llms-full.txt: every docs page concatenated into a single Markdown file.
- agents.txt: concise instructions for agents using these docs.
- api/search?q=YOUR_QUERY: deterministic docs search.
All files are generated fresh on every deploy.