Self-Hosted / Projects / LibreChat
LibreChat
StableOverview
Self-hosted ChatGPT alternative with multi-model access, agents, MCP, search, and team authentication.
LibreChat is a self-hosted ChatGPT alternative that brings multiple AI providers and conversation features into one web interface. It is designed for people who want a unified front end while keeping deployment, configuration, and user access under their control.
The important distinction is that LibreChat does not provide a model for free by itself. You connect supported providers or local endpoints, then operate the web app, database, configuration, and security boundary yourself.
Why people choose LibreChat
The official site lists model selection across providers such as Anthropic, AWS, OpenAI, Azure, and more, alongside agents, code interpretation, artifacts, message search, MCP, memory, web search, and enterprise authentication options.
That makes LibreChat a broader self-hosted AI workspace than a single local-model launcher. It can centralize cloud APIs and local or custom endpoints, but the resulting cost, privacy, and performance depend on the providers and tools you configure.
What the stack includes
The official Docker guide recommends Docker Compose, starts the app at port 3080, and uses a MongoDB-backed compose setup. The project also documents optional configuration through librechat.yaml and additional services or endpoints for search, memory, agents, and tools.
Authentication options include OAuth, LDAP, SAML, and two-factor authentication in the product documentation. These are useful for teams but should be treated as configuration and security work, not assumed to be enabled by default.
Requirements and tradeoffs
The supported baseline is Git plus Docker or Docker Desktop and a persistent compose deployment. The official docs warn that Apple Silicon Macs need a MongoDB override because the default MongoDB image requires AVX instructions.
LibreChat is a flexible integration layer, not a magic private AI box: API usage can still send data to external providers, local models need their own compute, and tool access, file handling, code execution, and web search increase the security surface.
Best fit
Choose LibreChat when you want one self-hosted interface for multiple AI providers, agents, tools, and team access. Consider Open WebUI or a simpler local-model front end when your priority is only chatting with local models and minimizing configuration.
Interface previews
Screenshots of the LibreChat interface.
Feature support
| Feature | Support |
|---|---|
| Multi-provider model accessOfficial pages document Anthropic, AWS, OpenAI, Azure, and other model connections. | Supported |
| Agents with file handling and API actionsAgents are a documented feature with tool and file workflows. | Supported |
| Code InterpreterThe official feature list documents code execution in multiple languages. | Supported |
| Artifacts for React, HTML, and MermaidArtifacts are listed as a supported feature. | Supported |
| MCP tool connectionsLibreChat documents Model Context Protocol support for tools and services. | Supported |
| Message, file, and code searchSearch is supported and may involve Meilisearch configuration. | Partial support |
| Persistent memoryMemory is available but needs appropriate configuration and provider support. | Partial support |
| OAuth, SAML, LDAP, and 2FADocumented authentication options require deliberate configuration and testing. | Partial support |
| Docker Compose deploymentThe official local installation path recommends Docker Compose. | Supported |
Community signals
- Stars
- 46.2k
- Forks
- 9.5k
- Open issues
- 999
- Last commit
- 6h ago
- Repo created
- Feb 2023
Includes open pull requests
Refreshed nightly from the GitHub API.
Alternatives
- Open WebUI
Local-model-focused web UI with a different integration and deployment emphasis.
- ChatGPT
Hosted alternative when infrastructure ownership is not required.
Questions
What is LibreChat used for?
LibreChat is used as a self-hosted interface for multiple AI models, conversations, agents, tools, code interpretation, search, and team authentication.
Is LibreChat a self-hosted ChatGPT alternative?
Yes. LibreChat provides a self-hosted ChatGPT-style interface, but you supply the model providers or local endpoints and remain responsible for the deployment.
Does LibreChat run AI models locally?
It can connect to local or custom endpoints, but it is primarily an interface and orchestration layer. Whether inference is local depends on the endpoints you configure.
What does LibreChat need for Docker deployment?
The official Docker guide lists Git and Docker or Docker Desktop as prerequisites and recommends Docker Compose. The standard compose setup includes MongoDB.
Does LibreChat need MongoDB?
The official Docker compose setup includes MongoDB for the application stack. Treat the database as a required deployment component unless you are following a different supported architecture.
Does LibreChat support OpenAI and Anthropic?
Yes. The official product site documents model selection across OpenAI, Anthropic, AWS, Azure, and other providers.
Is LibreChat private if it is self-hosted?
Self-hosting keeps the application and its configuration under your control, but prompts and files can still leave the server when you use external model or search providers. Review endpoint and tool settings before handling sensitive data.
Is LibreChat better than Open WebUI?
Neither is universally better. LibreChat is a strong fit for multi-provider access, agents, MCP, and team features; Open WebUI may be simpler when the main goal is a local-model chat interface.
support // the lab
Found this write-up useful?
If it saved you time or a rebuild, you can support more practical homelab guides.