Chatraw
Ultra-minimal AI chat UI: 30s deploy, no sign-up; OpenAI-compatible; RAG + vision + web parsing; plugins/adapters.
Last verified:
What is Chatraw?
ChatRaw is an ultra-minimal, fast, and lightweight AI chat UI designed for developers, AI hardware vendors, researchers, students, and enterprises who need a simple application to quickly demonstrate model capabilities. It deploys in 30 seconds via Docker with no complex configuration or heavy dependencies, offering a clean and fast AI chat experience. The interface is fully customizable with your own name, logo, subtitle, avatar, and theme colors, and supports bilingual English/Chinese UI with one-click switching.
Key features include universal support for any OpenAI-compatible API (Ollama, vLLM, LocalAI, LM Studio, etc.), native document parsing for PDF, DOCX, TXT, and MD files as chat context, and vision AI readiness with multimodal image understanding and auto-compression to WebP (~2MB). It supports thinking mode for reasoning models like DeepSeek-R1, Qwen, and o1 with collapsible thought process display. The UI is responsive and touch-friendly across desktop, tablet, and mobile devices, with one-click copy for AI responses (text only, no metadata), zero registration (settings auto-saved locally), and automatic API key rotation to bypass rate limits.
ChatRaw features a complete extensible plugin system with official plugins including Lightweight RAG Demo for knowledge base retrieval, Bocha Search and Tavily Search for web/AI search, Excel and CSV parsers, Enhanced Web Parsing (browser/Firecrawl/Jina), Multi-Model Manager, Markdown Renderer Plus with KaTeX math and Mermaid support, Context Compressor for automatic chat history summarization, and Toolbar Extension Demo. Plugins can add custom toolbar buttons with active/loading states, overflow menus, and fullscreen modals. The system includes complete development documentation, rich hook system with UI Extension API, custom settings UI, and one-click packaging for distribution.
The tool is extremely performant with approximately 60MB memory footprint, optimized binary embedding storage, and connection pooling for fast API calls. Performance scores are Desktop: Performance 100, Accessibility 100, Best Practices 100, SEO 100; Mobile: Performance 96, Accessibility 93, Best Practices 100, SEO 100. It supports unlimited chat, embedding, and reranking models with built-in endpoint validation and testing.
Chatraw pricing
Pricing model: Freemium
Completely free and open-source under MIT License. No paid plans, subscription tiers, or fees. Self-hosted deployment via Docker or from source. All features including plugin system are free. Docker images available on Docker Hub (massif01/chatraw) and GitHub Container Registry (ghcr.io/massif-01/chatraw) at no cost.
Chatraw pros
- Ultra lightweight ~60MB memory footprint
- 30-second Docker deployment
- Zero registration required
- Universal OpenAI-compatible API support
- Native PDF, DOCX, TXT, MD document parsing
- Vision AI with multimodal image understanding
- Thinking mode for reasoning models
- Automatic API key rotation for rate limits
- Responsive design for desktop, tablet, mobile
- Bilingual English and Chinese UI
- One-click copy AI responses (text only)
- Fully customizable branding and theme
- Complete plugin system with market
- 100 Lighthouse performance score on desktop
- Auto-saved local settings
- Context Compressor for history summarization
- Enhanced web parsing with multiple options
- Markdown Renderer Plus with KaTeX and Mermaid
- Excel and CSV parser plugins included
- Touch-friendly mobile UI
Chatraw cons
- Requires Docker or Python 3.12+ for deployment
- No built-in cloud hosting (self-hosted only)
- RAG moved to plugin in v2.0 (not core)
- Default theme is light (must change in settings)
- One-click copy excludes metadata
- Images auto-compressed to ~2MB WebP
- Plugin system requires installation steps
- LAN access needs host network mode
Frequently asked questions about Chatraw
What is ChatRaw?
ChatRaw is an ultra-minimal AI chat UI that deploys in 30 seconds with no sign-up. It is fast, lightweight (~60MB memory), and works with any OpenAI-compatible API including Ollama, vLLM, LocalAI, and LM Studio. It features document parsing, vision support, web page parsing, RAG, and a plugin system for extensibility.
How do I deploy ChatRaw?
Use Docker (recommended): run 'docker pull massif01/chatraw:latest' then 'docker run -d -p 51111:51111 -v chatraw-data:/app/data --name chatraw massif01/chatraw:latest'. Access at http://localhost:51111. Alternatively, clone the repo, install Python 3.12+ dependencies, and run 'python main.py' in the backend folder.
What APIs does ChatRaw support?
ChatRaw works with any OpenAI-compatible API including Ollama, vLLM, LocalAI, LM Studio, and official OpenAI APIs. You configure the API Base URL, Model ID, and API Key in Settings → Model Settings. It supports unlimited chat, embedding, and reranking models with automatic API key rotation.
Does ChatRaw support document uploads?
Yes, ChatRaw natively parses PDF, DOCX, TXT, and MD files as chat context. The AI can read and reference document content. You can also attach images for multimodal understanding, which are automatically compressed to WebP format (~2MB).
What plugins are available for ChatRaw?
Official plugins include: Lightweight RAG Demo (knowledge base retrieval), Bocha Search (web/AI search), Tavily Search (web search + AI answers), Excel Parser, CSV Parser, Enhanced Web Parsing (browser/Firecrawl/Jina), Multi-Model Manager, Markdown Renderer Plus (KaTeX, Mermaid, code copy), Context Compressor, and Toolbar Extension Demo.
How do I install plugins?
Click the Plugins button in the bottom-left corner, browse the Plugin Market tab, click Install on any plugin, then enable it in the Installed tab. For Context Compressor, enable the plugin and configure autoCompress setting for automatic or manual compression via the input toolbar button.
Can I customize the ChatRaw interface?
Yes, in Settings → Interface you can customize the application name, logo, subtitle, user and AI avatars, and theme colors. You can also switch between light and dark mode, and toggle between English and Chinese UI with one click.
Does ChatRaw support reasoning models?
Yes, ChatRaw supports thinking mode for reasoning models like DeepSeek-R1, Qwen, and o1. The thought process is displayed in a collapsible format so you can view or hide the AI's reasoning steps.
How do I update ChatRaw?
For Docker (docker run): stop and remove the container, run 'docker pull massif01/chatraw:latest', then start again with the same volume to preserve data. For Docker Compose: run 'git pull origin main', 'docker compose pull', then 'docker compose up -d'. For source: 'git pull origin main', then 'pip install -r requirements.txt --upgrade' and run main.py.
Is ChatRaw free to use?
Yes, ChatRaw is completely free and open-source under the MIT License. There are no paid plans, subscriptions, or fees. All features including the plugin system are free. Docker images are available on Docker Hub and GitHub Container Registry at no cost.