Witsy

Normal
GitHub TypeScript AGPL-3.0

Description

Witsy: desktop AI assistant and universal MCP client that connects to multiple AI models and tool services via the Model Context Protocol.

Key Features

  • Universal MCP client supporting any LLM provider including OpenAI, Anthropic, Google, Ollama, and more
  • Prompt Anywhere feature generates content directly in any application via global keyboard shortcut
  • AI Commands provide quick text-processing helpers (summarize, translate, rewrite) on selected text
  • RAG chat with local documents using OpenAI embeddings or Ollama embedding models
  • Multi-modal support: text-to-image, text-to-video, speech-to-text, and text-to-speech
  • Expert prompts and long-term memory plugin for specialized and context-aware conversations

Use Cases

💡 Use any LLM provider with MCP servers through a single unified desktop interface
💡 Generate and insert AI content directly into any application without copy-paste
💡 Chat with local documents for research and knowledge retrieval tasks
💡 Transcribe audio and dictate text using multiple speech-to-text backends
💡 Create images, videos, and voiceovers through integrated multi-modal providers

Categories

Quick Start

Download from the GitHub releases page or install via `brew install --cask witsy`. Configure your API keys for the desired LLM providers in Settings. Use Shift+Control+Space to trigger Prompt Anywhere in any application.

Related Projects