Witsy

Witsy is a desktop AI assistant for people who want to chat with models, work with documents and write across applications. You can choose local models or cloud providers, with feature availability depending on the backend.

• Model chat and tools: Use Ollama, LM Studio, OpenAI, Anthropic, Gemini or OpenAI-compatible providers, and connect MCP servers so models can use external tools.
• Document chat: Ask questions about local files, with relevant passages sent to your chosen model and embeddings handled by Ollama or OpenAI.
• Writing across apps: Edit drafts in the scratchpad, insert generated text with Prompt Anywhere and apply custom AI commands to highlighted text in other applications.
• Voice and media: Transcribe microphone audio with local Whisper or cloud services, use voice conversations and read-aloud features, and analyze images with compatible models; image creation supports Stable Diffusion WebUI and cloud backends, while video creation uses cloud providers.
• Automation: Chat from a terminal while the desktop app runs, connect other applications through a local HTTP API and trigger agents through webhooks.

Witsy runs on macOS, Windows and Linux and is open source under AGPL-3.0. Ollama runs models on your machine for free. Cloud providers require your own API keys and receive requests sent to them. Conversation history stays locally, but document passages go to the selected model, and OpenAI embeddings send document content to an external service.