
Open-source desktop AI assistant with 12 modes, multi-model support, and plugin system
PyGPT is a free, open-source desktop AI assistant for Windows, macOS, and Linux that supports 12 operational modes including chat, vision, agents, autonomous mode, image/video generation, voice control, and computer use. It connects to multiple AI providers — OpenAI (GPT-5, GPT-4, o1, o3), Google Gemini, Anthropic Claude, xAI Grok, DeepSeek, Perplexity, Mistral, and local models via Ollama. PyGPT includes built-in RAG with vector database support for chatting with documents (PDF, CSV, DOCX, and more), 20+ plugins, Python code execution, system command integration, speech synthesis/recognition, web search, long-term memory, and accessibility features for users with disabilities.
Chat, Chat with Files, Realtime Audio, Research (Perplexity), Completion, Image/Video Generation, Vision, Assistants, Experts, Computer Use, Agents, and Autonomous Mode
Works with OpenAI GPT-5/4/o1/o3, Google Gemini, Anthropic Claude, xAI Grok, DeepSeek, Perplexity, Mistral, and local models via Ollama
Chat with your data — supports PDF, CSV, DOCX, HTML, JSON, EPUB, XLSX, XML, webpages, GitHub repos, video, audio, and images via LlamaIndex
Generate and run Python code, execute system commands, and manage file operations directly from the assistant
Speech synthesis via Azure, Google, ElevenLabs, and OpenAI TTS, plus speech recognition for hands-free operation
Create images with DALL-E and videos with Sora 2 and Veo 3 directly from the chat interface
20+ built-in plugins with extensible architecture for adding custom functionality and tools
Customizable keyboard shortcuts, voice control, and on-screen action translation to audio for users with disabilities