LibreChat

Self-hosted multi-provider chat UI with agents, MCP tools, code execution and multi-user login.

Open SourceSelf Hosted
0.0 (0)

About

Organizations wanting one self-hosted chat front end for many vendors use LibreChat to put OpenAI, Azure OpenAI, Anthropic, Gemini and Vertex AI, AWS Bedrock, and any OpenAI-compatible endpoint such as Ollama, OpenRouter, Groq or DeepSeek behind a ChatGPT-style interface with per-user accounts. Beyond chat it offers no-code agents with tool calling, MCP servers, file search, subagents and SKILL.md bundles, a sandboxed code interpreter, web search with reranking, React, HTML and Mermaid artifacts, image generation, speech input and output, conversation forking, presets, and full-text message search. Sign-in options include email, OAuth2, OpenID and LDAP, with roles, groups, token spend controls and a browser admin panel. The usual install is Docker Compose, which starts the Node.js app with MongoDB, Meilisearch, a RAG API and a vector database; an npm route needs Node 24 and your own MongoDB, and a Helm chart targets Kubernetes. Code is MIT, so commercial deployments are allowed; the project website says it is joining ClickHouse. A strong pick for teams wanting control over data, providers and users.

Should you use LibreChat?

Pick it when

Pick LibreChat when an organization wants a self-hosted ChatGPT replacement for many users, with login, mixed providers like OpenAI, Azure, Anthropic, and Google in one place, plus agents, MCP, code interpreter, and conversation search.

Look elsewhere when

Skip it if you are one person or mainly run local models: it is a Docker stack with more moving parts. Open WebUI is easier for Ollama-centric setups, and Chatbox or NextChat suit a single user with API keys.

Alternatives to LibreChat

  • Open WebUI

    Also multi-user, with a stronger focus on Ollama and local models plus built-in RAG; easier when your models run on your own hardware.

  • Lobe Chat

    Polished agent workspace with marketplace and scheduled runs, but its community license limits distributing modified versions, unlike LibreChat's MIT.

  • Hugging Face Chat UI

    Leaner front end for OpenAI-compatible endpoints with routing, lighter to run, but without LibreChat's native provider connectors.

  • big-AGI

    Personal workspace with Beam multi-model comparison and client-side data, quicker to start, but not a multi-user server with accounts.

Reviews (0)

Leave a Review

No reviews yet. Be the first to review!