Open WebUI

Overview

Open WebUI is a self-hosted AI platform designed to be a home for AI. It is extensible, feature-rich, and user-friendly, with support for Ollama and OpenAI-compatible APIs. It can run entirely offline and provides a provider-agnostic interface for both local and cloud-based models.

Key Features

• Effortless Setup: Install via pip, uv, Docker, or Kubernetes with official images for Ollama and CUDA.
• Broad Model & API Integration: Connect any OpenAI-compatible API alongside local Ollama models, including LMStudio, GroqCloud, Mistral, OpenRouter, vLLM, and more.
• Granular RBAC & User Groups: Administrators define detailed roles, groups, and permissions for tailored access.
• Plugin Support: Extend with Filters, Actions, Pipes, Tools, and Skills, and connect external services via MCP, MCPO, and OpenAPI tool servers.
• Local RAG Integration: Retrieval Augmented Generation with 9 vector databases and multiple content-extraction engines, supporting hybrid search and full-context mode.
• Multi-Model Conversations: Engage several models at once to harness their individual strengths in parallel.

Use Cases

• Self-hosted AI chat interface for local and cloud models.
• Team collaboration with role-based access and shared channels.
• Enterprise deployment with SSO, SCIM, and production observability.
• Building custom agents with plugins, tools, and persistent memory.

Getting Started

• Install via pip: pip install open-webui, then run open-webui serve. For Docker, use: docker run -d -p 3000:8080 -v open-webui:/app/backend/data --name open-webui --restart always ghcr.io/open-webui/open-webui:main. Access at http://localhost:3000 (or localhost:8080 with

Deployment & Requirements

• Python 3.11 is required for pip installation. Docker is recommended for container deployment, with optional CUDA support requiring the Nvidia CUDA container toolkit. Database options include SQLite (with optional encryption) or PostgreSQL, and file storage can be local or on S3,

Before You Adopt

• License: Other. Review its terms before using, modifying, or distributing the project.
• The :dev branch contains unstable features and may have bugs.
• Docker installations must mount the data volume to prevent data loss.
• Connection issues may require using --network=host to reach Ollama.