mirror of
https://github.com/danny-avila/LibreChat.git
synced 2026-09-01 11:33:44 +00:00
Enhanced ChatGPT Clone: Features Agents, MCP, DeepSeek, Anthropic, AWS, OpenAI, Responses API, Azure, Groq, o1, GPT-5, Mistral, OpenRouter, Vertex AI, Gemini, Artifacts, AI model switching, message search, Code Interpreter, langchain, DALL-E-3, OpenAPI Actions, Functions, Secure Multi-User Auth, Presets, open-source for self-hosting. Active.
https://librechat.ai/
aianthropicartifactsawsazurechatgptchatgpt-cloneclaudeclonedeepseekgeminigooglegpt-5librechatmcpo1openairesponses-apivisionwebui
2f50e38217
* 🖼️ fix: Keep Composer Send Enabled When an Attachment Stalls The composer's send button is gated on `hasIncompleteFiles(files)`, so any attachment that can never reach `progress: 1` reads as "still uploading" and disables send for the rest of the session — draft text intact, no error, no way out but removing the chip or reloading. Two paths could park an attachment there: - `loadImage` starts the upload from `img.onload` and had no `onerror`, so an image the browser refuses to decode (unsupported codec, truncated bytes, a revoked object URL) never uploaded at all and stranded the file at `progress: 0.2`. Drop the file and surface the error instead. - Upload completion reconciled against `temp_file_id`, the server's echo of the id the request was sent with, while every client-side handle for that upload — file map key, delayed-toast timer, recovery callbacks — is keyed by the id the client owns. A mismatch applied the completion update to a key that does not exist, leaving the attachment at `progress: 0.9`. Covered by unit regressions in the file-handling suite and a composer-level spec that drives a real upload through `ChatForm`, plus a render-bound guard on typing (react-scan measures one ChatForm render per keystroke in a browser; the guard fails on a multiplier). Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Cq3tev2rPbc2pWyVJwnh6u * 🧹 fix: Stop the Draft Restore From Clobbering Live Composer Attachments `restoreFiles` runs on every `QueryKeys.files` write — an upload landing, an SSE attachment mid-run — not just on a conversation swap, and it was written as if the draft were always the whole truth: - An empty draft cleared the composer outright. On the swap path that is redundant (the effect already clears explicitly one line earlier); on the cache path an empty draft only means the draft write has not caught up, so clearing there discards an attachment the user just added — and with no text typed, the send button has nothing left to submit. Restoring now only adds. - A match replaced the composer's entry with the persisted record, dropping the local `File`, the blob preview the chip renders from (`FileRow` falls back to refetching `filepath`), and the tool resource the upload was staged under, and stamping `attached: true` so removing a chip the composer still owns leaves the file orphaned server-side. It now layers the record over the live entry and leaves `attached` to files actually adopted from a draft. Confirmed against a real browser run: the entry is at `progress: 0.9` when this restore fires, so it — not the upload's own completion — is what was re-enabling send. react-scan render counts are unchanged (typing 20 keystrokes: 111 renders, ChatForm=20; attaching an image: 1373, FileRow=6). Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Cq3tev2rPbc2pWyVJwnh6u * 🔗 fix: Keep an Attachment's Stored Temporary Id Equal to Its Map Key Two follow-ups from review on the upload reconciliation. Completion stored the server's `temp_file_id` echo in the entry's value while keying the map by the id the request was sent with. `useFileDeletion` deletes map entries by the value's own `file_id` and `temp_file_id`, so where the two disagreed — the exact case the reconciliation exists to tolerate — Remove would delete the file server-side and leave the chip behind, and the draft restore could not correlate its saved key with the cached record. Store the request id. A refused image decode also left its `uploadScope.recent` reservation behind: reservations are released by the render that observes the file in the shared state, which a decode failing before that render never reaches, and once the file is deleted no later render can either. The ghost is merged into every later batch's validation, so re-picking the same file reads as a duplicate and its size keeps counting against the composer's limits. Both covered; both new guards fail without their fix. Also sorts the composer spec's imports, which the static-checks import-order gate flagged. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Cq3tev2rPbc2pWyVJwnh6u * 🧷 fix: Normalize an Upload's Temporary Id at the Cache Boundary The composer keys its file map — and the draft it saves — by the `file_id` the upload request was sent with; `temp_file_id` is only the server's echo of that id. The previous commit reconciled the composer's own entry against the request id but left the record the mutation inserts into `QueryKeys.files` carrying the raw echo, and `restoreFiles` can only correlate a saved draft id by matching a cached record's `file_id` or `temp_file_id`. Where the echo disagreed the draft matched neither, so the attachment was silently dropped on the next conversation switch or reload — the same class of loss, one layer further out. Normalize once where the response enters client state, and hand the normalized record to the mutation's callers, so the cache, the composer entry and the draft all agree on one id. An agreeing response is passed through untouched. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Cq3tev2rPbc2pWyVJwnh6u --------- Co-authored-by: Claude <noreply@anthropic.com> |
||
|---|---|---|
| .claude/skills | ||
| .devcontainer | ||
| .github | ||
| .husky | ||
| .vscode | ||
| api | ||
| client | ||
| config | ||
| e2e | ||
| helm | ||
| otel/langfuse-fanout | ||
| packages | ||
| redis-config | ||
| scripts | ||
| search | ||
| skill | ||
| src/tests | ||
| utils | ||
| .dockerignore | ||
| .env.example | ||
| .gitattributes | ||
| .gitignore | ||
| .nvmrc | ||
| .prettierrc | ||
| AGENTS.md | ||
| bun.lock | ||
| CLAUDE.md | ||
| CONTEXT.md | ||
| deploy-compose.langfuse-fanout.yml | ||
| deploy-compose.yml | ||
| docker-compose.langfuse-fanout.yml | ||
| docker-compose.override.yml.example | ||
| docker-compose.yml | ||
| Dockerfile | ||
| Dockerfile.multi | ||
| eslint.config.mjs | ||
| librechat.example.yaml | ||
| LICENSE | ||
| package-lock.json | ||
| package.json | ||
| rag.yml | ||
| README.md | ||
| README.zh.md | ||
| tool-intent-spec.md | ||
| turbo.json | ||
LibreChat
English · 中文
🚀 What's New in v0.8.8-rc1
- Agent run control: Interrupt or steer an Agent mid-run, queue follow-up messages, and reclaim, edit, or escalate pending steers.
- Human-in-the-loop Agents: Agents stream question progress, ask up to four related questions in one form, pause for input or tool approval, and resume.
- Unified Agent Builder: A redesigned Tools marketplace brings together Skills, MCP, Code Interpreter, orchestration, Programmatic Tool Calling, model-spec controls, and per-tool background and intent settings.
- Readable Agent activity: Generated activity-group headers, parent phase summaries, and live tool intent labels make long reasoning and tool runs easier to scan.
- Code Interpreter workflows: Code and shell tools can run in the background, sandbox images return as viewable artifacts, and highly experimental stateful sessions can reuse prewarmed conversation workspaces.
- Agent extensibility: Experimental Agent Plugins can bundle deployment Skills, MCP servers, and opt-in command hooks, while explicit subagents initialize only when selected.
- Memory, context, and identity: Agents can manage memory with optional per-agent isolation, expose support contacts safely, and show a more faithful Context Usage gauge.
- Sharing and files: Shared conversations show a badge and update at a stable URL, while signed-in viewers can continue them as personal copies.
- Artifact workflows: Open previews fullscreen, work with PowerPoint
.potxtemplates across upload, search, and code execution, upload shell scripts across common MIME variants, export Mermaid diagrams as SVG or PNG, and download original Office files from the artifact panel. - Models and reasoning: Added GPT-5.6 with Responses API reasoning controls, Claude Opus 5 and Sonnet 5, Gemini 3.7 and 3.6 Flash, and Gemini 3.5 Flash-Lite.
- Langfuse observability: Configure encrypted Langfuse connections in-app, let authorized admins open sampled sessions directly, optionally fan out traces by tenant, and suppress central export per run.
- Administration and security: Delegate config sections, encrypt registered secrets, enforce SSRF checks for speech, OCR, and web tools, and generate unique temporary credentials when secrets are blank.
- Messages and navigation: Right-aligned user turns, unified multi-part editing, full-message copy, a dock-style message rail, virtualized search, smooth streaming, and faster Agent startup.
- Streaming and tool reliability: Adaptive provider smoothing, Redis delta batching, dynamic MCP tool refresh, parsed MCP response media types, runtime OAuth recovery, and Agent stream circuit breakers improve long-running workflows.
- Deployment and reliability: Added configurable HTTP timeouts, Amazon DocumentDB 5.0+ support, low-noise Redis and browser observability, and a rolling-upgrade-safe generation protocol.
Read the full v0.8.8-rc1 changelog.
✨ Features
-
🖥️ UI & Experience inspired by ChatGPT with enhanced design and features
-
🤖 AI Model Selection:
- Anthropic (Claude), AWS Bedrock, OpenAI, Azure OpenAI, Google, Vertex AI, OpenAI Responses API (incl. Azure)
- Custom Endpoints: Use any OpenAI-compatible API with LibreChat, no proxy required
- Compatible with Local & Remote AI Providers:
- Ollama, groq, Cohere, Mistral AI, Apple MLX, koboldcpp, together.ai,
- OpenRouter, Helicone, Perplexity, ShuttleAI, Deepseek, Qwen, and more
-
- Secure, Sandboxed Execution in Python, Node.js (JS/TS), Go, C/C++, Java, PHP, Rust, and Fortran
- Seamless File Handling: Upload, process, and download files directly
- No Privacy Concerns: Fully isolated and secure execution
- Open-Source & Self-Hostable: powered by ClickHouse/code-interpreter
-
🔦 Agents & Tools Integration:
- LibreChat Agents:
- No-Code Custom Assistants: Build specialized, AI-driven helpers
- Agent Marketplace: Discover and deploy community-built agents
- Collaborative Sharing: Share agents with specific users and groups
- Flexible & Extensible: Use MCP Servers, tools, file search, code execution, and more
- Skills: Create reusable
SKILL.mdinstruction bundles for manual, automatic, or always-on agent workflows - Agent Plugins: Experimentally bundle deployment Skills and MCP servers into startup-loaded packages
- Subagents: Delegate focused work to isolated child agent runs with their own context windows
- Compatible with Custom Endpoints, OpenAI, Azure, Anthropic, AWS Bedrock, Google, Vertex AI, Responses API, and more
- Model Context Protocol (MCP) Support for Tools
- LibreChat Agents:
-
🔍 Web Search:
- Search the internet and retrieve relevant information to enhance your AI context
- Combines search providers, content scrapers, and result rerankers for optimal results
- Customizable Jina Reranking: Configure custom Jina API URLs for reranking services
- Learn More →
-
🪄 Generative UI with Code Artifacts:
- Code Artifacts create React, HTML, and Mermaid content directly in chat
- Open previews fullscreen and export Mermaid diagrams as SVG or PNG
-
🎨 Image Generation & Editing
- Text-to-image and image-to-image with GPT-Image-1
- Text-to-image with DALL-E (3/2), Stable Diffusion, Flux, or any MCP server
- Produce stunning visuals from prompts or refine existing images with a single instruction
-
💾 Presets & Context Management:
- Create, Save, & Share Custom Presets
- Switch between AI Endpoints and Presets mid-chat
- Edit, Resubmit, and Continue Messages with Conversation branching
- Create and share prompts with specific users and groups
- Fork Messages & Conversations for Advanced Context control
-
💬 Multimodal & File Interactions:
- Upload and analyze images with Claude 3, GPT-4.5, GPT-4o, o1, Llama-Vision, and Gemini 📸
- Chat with Files using Custom Endpoints, OpenAI, Azure, Anthropic, AWS Bedrock, & Google 🗃️
-
🌎 Multilingual UI:
- English, 中文 (简体), 中文 (繁體), العربية, Deutsch, Español, Français, Italiano
- Polski, Português (PT), Português (BR), Русский, 日本語, Svenska, 한국어, Tiếng Việt
- Türkçe, Nederlands, עברית, Català, Čeština, Dansk, Eesti, فارسی
- Suomi, Magyar, Հայերեն, Bahasa Indonesia, ქართული, Latviešu, ไทย, ئۇيغۇرچە
-
🧠 Reasoning UI:
- Dynamic Reasoning UI for Chain-of-Thought/Reasoning AI models like DeepSeek-R1
-
🎨 Customizable Interface:
- Customizable Dropdown & Interface that adapts to both power users and newcomers
-
- Never lose a response: AI responses automatically reconnect and resume if your connection drops
- Multi-Tab & Multi-Device Sync: Open the same chat in multiple tabs or pick up on another device
- Production-Ready: Works from single-server setups to horizontally scaled deployments with Redis
-
🗣️ Speech & Audio:
- Chat hands-free with Speech-to-Text and Text-to-Speech
- Automatically send and play Audio
- Supports OpenAI, Azure OpenAI, and Elevenlabs
-
📥 Import & Export Conversations:
- Import Conversations from LibreChat, ChatGPT, Chatbot UI
- Export conversations as screenshots, markdown, text, json
-
🔍 Search & Discovery:
- Search all messages/conversations
-
👥 Multi-User & Secure Access:
- Multi-User, Secure Authentication with OAuth2, LDAP, & Email Login Support
- Built-in Moderation, and Token spend tools
-
🎛️ Admin Panel:
- Browser-based UI to manage users, groups, roles, and configuration overrides
- Edit settings and per-role/group permissions live, without redeploying
- Bundled with the Docker Compose stacks for one-command setup
-
⚙️ Configuration & Deployment:
- Configure Proxy, Reverse Proxy, Docker, & many Deployment options
- Use S3 with CloudFront for stable media links, edge delivery, signed cookies, and secured downloads
- Use completely local or deploy on the cloud
-
📖 Open-Source & Community:
- Completely Open-Source & Built in Public
- Community-driven development, support, and feedback
For a thorough review of our features, see our docs here 📚
🪶 All-In-One AI Conversations with LibreChat
LibreChat is a self-hosted AI chat platform that unifies all major AI providers in a single, privacy-focused interface.
Beyond chat, LibreChat provides AI Agents, Model Context Protocol (MCP) support, Artifacts, Code Interpreter, custom actions, conversation search, and enterprise-ready multi-user authentication.
Open source, actively developed, and built for anyone who values control over their AI infrastructure.
🌐 Resources
GitHub Repo:
- RAG API: github.com/danny-avila/rag_api
- Website: github.com/LibreChat-AI/librechat.ai
Other:
- Website: librechat.ai
- Documentation: librechat.ai/docs
- Blog: librechat.ai/blog
📝 Changelog
Keep up with the latest updates by visiting the releases page and notes:
⚠️ Please consult the changelog for breaking changes before updating.
⭐ Star History
✨ Contributions
Contributions, suggestions, bug reports and fixes are welcome!
For new features, components, or extensions, please open an issue and discuss before sending a PR.
If you'd like to help translate LibreChat into your language, we'd love your contribution! Improving our translations not only makes LibreChat more accessible to users around the world but also enhances the overall user experience. Please check out our Translation Guide.
💖 This project exists in its current state thanks to all the people who contribute
🎉 Special Thanks
We thank Locize for their translation management tools that support multiple languages in LibreChat.