Sider
AssistantsRead pages with Chat, handle browser tasks with Hand and customize your browsing experience with Code. Sider adds an AI workspace to the sites and documents you already use.
Chat beside the sourceHand browser tasks
AI WORKSPACE / MODELS · AGENTS · PROJECTS
LibreChat
A free self-hosted AI chat platform that combines multiple providers, reusable agents, project conversations and optional file or tool workflows in one web interface.
Official LibreChat website image showing model choices, projects and the chat workspaceExplore the image ↓
What you can do with it
What this tool is for
LibreChat provides a common web interface for conversations across several AI providers. A self-hosted deployment can organize models, reusable agent instructions, files and projects without tying every workflow to a single provider’s website. It suits people who are comfortable managing an application and want control over their interface, access arrangements and stored conversations.
It is a web application, not a native desktop installer that automatically brings a local model with it. The official documentation recommends a Docker-based starting route and also describes other installation choices. Models and optional services must be configured separately. Begin with a working chat connection before adding agents, document retrieval or code execution.
01
Compare two approaches
02
A small team chat portal
03
A reusable research assistant
04
Prototype an interactive explanation
A closer look at the product
Official product image. Click to inspect the details.
Product interface and examples
Key features, practical tasks
Capabilities in practice
- Provider choices: Connect supported model providers or custom compatible endpoints. Capabilities vary by endpoint, so check whether the selected model supports the files, images and tools required by a task.
- Reusable agents: Combine instructions, model settings and selected tools for a repeatable job. An agent should have a defined purpose and limited access rather than a large collection of unrelated capabilities.
- Projects and context: Group related conversations and project material so repeated work has a consistent starting point. Keep unrelated clients or subjects separate when their documents and instructions should not mix.
- Conversation forks: Branch from a message into a separate chat and choose how much history to copy. This is useful for exploring another draft or interpretation without changing the original conversation.
- Files and retrieval: The project supports file-oriented workflows and a RAG API ecosystem. Distinguish providing a whole file as text from retrieving selected passages, and check what actually reaches the model.
- Artifacts and tools: Preview supported generated content and connect MCP or other tools when needed. Code generation, code display and actual code execution are different actions with different setup and cost implications.
- Conversation portability: The project documents conversation import and export, including readable text-oriented formats. Preserve important source context and reviewed deliverables instead of relying only on an attractive chat screenshot.
Get started, step by step
- 01
Install the application
- 02
Connect one model
- 03
Define a small workspace or agent
- 04
Check and preserve a useful result
A practical first workflow
A practical first workflow
- Install the application. Follow the current quick-start route for your platform, prepare persistent storage and confirm the web interface loads. Read release notes before upgrading an existing deployment.
- Connect one model. Add the intended provider or local endpoint and verify ordinary text chat. This isolates account and connection problems from later agent or file-workflow issues.
- Define a small workspace or agent. Set one purpose, clear instructions and only the tools needed for the task. Add a short reference source rather than a large uncurated archive.
- Check and preserve a useful result. Try an alternate branch, inspect supporting material and export or save the reviewed output. Keep a recoverable deployment configuration for future changes.
Example prompt or task: Draft a concise explanation from the supplied notes. Keep facts, interpretation and open questions separate. Do not contact external services or change files; identify what further evidence would be needed before taking action.
This editorial example is appropriate for a first bounded agent. Add action tools only when the workflow actually needs them and the intended permissions are clear.
Find the right plan for you
Free self-hosting
Explore official plans ↗Free software Models, tools and hosting are separate
The free application does not bundle unlimited inference or every hosted execution service.
What is free and what can cost extra
LibreChat’s application is MIT-licensed and can be self-hosted without a software subscription. You still provide hosting, storage and model access. Local models can avoid per-request external inference billing, while paid API providers charge according to their own accounts and terms.
Optional services, including some code-execution routes, search providers and other tools, may have separate costs. The public demo is an evaluation surface, not a promise of permanent free production hosting or private data storage. The free route listed here is your own deployment.
A few things worth knowing
Frequently asked questions
Is LibreChat free?
The MIT-licensed application provides a free self-hosted route. Infrastructure, model providers and optional external services remain separate.
Is it a native desktop AI app?
No. The official documentation describes a self-hosted web application. Running it locally still means setting up that application and its dependencies.
Does it include free access to every listed model?
No. A supported provider integration is not an entitlement to its models. Configure an account or compatible local service with the necessary access.
What is a conversation fork?
It copies selected history into a new conversation so you can explore a different direction without altering the original thread.
Are code artifacts the same as code execution?
No. Displaying or previewing generated content differs from running code through an execution service. Check the configured route, permissions and costs.
Can I use the public demo for private files?
The official docs frame the demo as evaluation-only and advise against using it for private or important retained data. Use an appropriately configured deployment for your own work.
Will the latest docs exactly match my installed version?
Not always. The documentation can describe development features. Match the guide and feature expectations to the release you actually run.
The details behind your decision
Explore platforms, inputs and outputs, licensing, and access requirements.
Four useful starting projects
Four useful starting projects
- Compare two approaches: Fork an established conversation to try a different prompt or model. Compare the resulting drafts while retaining the original assumptions and source material.
- A small team chat portal: Provide a consistent interface to approved model services. Configure accounts and access deliberately rather than sharing a single provider credential in ordinary chat messages.
- A reusable research assistant: Give an agent a narrow research task and a controlled set of files or tools. Ask it to separate retrieved evidence from its own suggested interpretation.
- Prototype an interactive explanation: Use supported artifacts to inspect generated HTML, diagrams or interface ideas. Treat the preview as a prototype and review its behavior before publishing or integrating it elsewhere.
Platforms, inputs, outputs and setup
Platforms, inputs, outputs and setup
The official docs explicitly describe LibreChat as a self-hosted web app rather than a native Windows application or Linux AppImage. Docker is a common local installation path, including Docker Desktop on Windows; other documented deployment routes serve different operational needs.
The docs can display a latest-development version, while your installed release may be older. Check the chosen release’s compatibility before planning around a new agent or administration feature. Model credentials, databases, uploaded files and application settings also need a deliberate backup and access arrangement.
Data, permissions and licensing
Data, permissions and licensing
Self-hosting lets you manage the application and its stored data. It does not prevent a configured external model or tool from receiving the content required to answer or act. Review file uploads, retrieval, search and tool calls as separate data paths.
Restrict administrative access and keep provider credentials out of shared prompts or exported examples. MIT covers the application code; model terms, external service conditions and rights to uploaded documents are independent.
Sources, documentation and help
Official sources and review date
Official sources and review date
Reviewed October 3, 2026. Product facts come from the official sources below. Suggested projects, prompts and review methods are FindGoodAI editorial guidance, not measured performance results.
Your experience can help someone choose
COMMUNITY
Sign in to rate and comment
Tried it? Share your experience
Loading experiences…
Manage my comments
Community and moderation
Saved tools are private. Approved comments are public; edits return to moderation. Comment counts include approved comments and replies. Each person has one rating. New ratings are automatically approved; withdrawn, unapproved or invalid ratings do not count. Share real experiences and avoid spam, private information or personal attacks.
For moderation appeals or data requests: support@findgoodai.com
NEXT STEP
Explore LibreChat ↗Start with one model and a clear workspace
Bring a real task and see how it fits the way you work.
Jump to a section
User experiences