TypingMind

TypingMind is an AI chat frontend for people who want to use different model providers in one workspace. It brings conversations, project materials and custom AI agents into a shared interface.

• Model comparison: Chat with OpenAI, Claude or Gemini, connect custom model endpoints, and compare responses from multiple models in parallel.
• Projects and chat history: Group conversations and knowledge bases into project folders, find chats with search and tags, and fork a conversation to explore another direction.
• Documents and knowledge: Ask questions about uploaded PDFs and other files, or give models access to material from GitHub, Notion and Google Drive through retrieval.
• Custom agents and tools: Give agents their own instructions, models and reference files, connect external tools through plugins or MCP, and link agents in prompt chains for multi-step tasks.
• Writing and coding: Work with AI in a Canvas editor and create editable artifacts, including documents and code, alongside rendered charts and diagrams.

You can use TypingMind through its website or deploy a static copy on your own server after purchase. The proprietary self-host package contains compiled code; it doesn't include the full source. The frontend connects to model services rather than running models itself. Cloud providers require your own API credentials and receive the requests you send to them.

Chats and API keys stay in your browser's local storage by default. Optional password encryption protects API keys, but it doesn't encrypt chat messages or prompts. Cross-device sync uses TypingMind Cloud, a separate, optional cloud storage service that requires an account; you can also export and import chats.