ChatFrame is a cross-platform desktop chatbot that unifies access to multiple LLM providers while offering local RAG for files and support for the Model Context Protocol (MCP).
What is ChatFrame?
ChatFrame is a desktop application that lets you chat with multiple large language model (LLM) providers — including OpenAI, DeepSeek, Anthropic, Groq, xAI, and Moonshot — through a single interface. It takes text and image inputs and produces text responses, with optional artifact rendering (React components, HTML, SVG, code snippets). The app runs on macOS (Apple Silicon and Intel) and Windows (x86_64). It is built using Tauri for the desktop framework and the Vercel AI SDK for LLM interactions.
Key Features
- Unified LLM Access — Interact with OpenAI, Anthropic, GoogleAIStudio, Qwen, and more without switching apps.
- MCP Support — Integrate custom tools via SSE, Streamable HTTP, or Stdio MCP servers (e.g., a Postgres MCP server with local runtime).
- Local RAG — Upload PDFs, text files, markdown, or code files into workspaces; build vector indexes locally for instant context-aware chat.
- Multimodal Chat — Send text and images, and render interactive artifacts like React, HTML, SVG, or code in real time.
- Privacy-Focused — No data is uploaded to external servers; all processing stays on your machine.
- Customizable Providers — Add API keys for supported providers or configure custom OpenAI-compatible endpoints.
- Cross-Platform — Available for macOS (Apple Silicon and Intel) and Windows (x86_64).
- Proxy & Shortcuts — Route all LLM API requests through a proxy; use keyboard shortcuts (⌘N new chat, ⌘B toggle sidebar).
Who is it for?
- Developers who want to test multiple LLMs or integrate MCP-based tools without leaving a chat interface.
- Researchers and students who need to query local documents (PDFs, code) using RAG while maintaining data privacy.
- AI enthusiasts who prefer a single desktop app over browser tabs for managing multiple AI service subscriptions.
What can you do with ChatFrame?
- Unified chat: Ask questions to OpenAI, Anthropic, or any configured provider in one conversation thread.
- Local document Q&A: Drop a PDF or code folder into a workspace, build a vector index, then ask questions about its contents offline.
- Tool integration: Configure an MCP server (e.g., Postgres) and instruct the LLM to query your database through natural language.
- Multimodal interactions: Share images to describe them or generate code artifacts that render inline.
How does ChatFrame work?
- Download the app from chatframe.co.
- Add your API keys via the Providers section and verify the configuration.
- Start chatting by clicking the Chat button. Optionally, create a workspace and upload files for RAG, or configure an MCP server for custom tools.
