

PyGPT
β Text source-checked Nov 21, 2026Open-source desktop AI assistant for Windows, Linux and macOS with agents, RAG, Python tools, MCP, vision, image/video generation and Ollama.
Source-checked against the vendor's site Β· rebuilt from the vendor's own pagesProfile updated Link checked How we rate
In short
PyGPT is a free, open-source desktop AI assistant for Windows, Linux, and macOS. It supports multiple models, RAG, agents, and local Ollama integration.
PyGPT is an advanced, open-source desktop AI assistant. The application runs on Windows, Linux, and macOS and is written in Python. PyGPT includes work modes for Chat, Agents, Realtime audio, Research, Completion, Image and Video generation, Computer use, Experts, and Autonomous modes.
The assistant supports models from OpenAI, Google, Anthropic, xAI, DeepSeek, Perplexity, HuggingFace, OpenAI-compatible APIs, and local Ollama models. PyGPT integrates LlamaIndex RAG for processing files, webpages, external sources, media, and conversation history. The app includes built-in Python and OS tools with support for IPython, system commands, a built-in sandbox, and Docker. PyGPT provides Model Context Protocol support with a built-in Connectors manager.
The application uses the user's own provider API credentials, while local models can run without external API keys. PyGPT is free and open source, with source code available on GitHub. The software is provided under the MIT License.
According to PyGPT: PyGPT is an advanced, Open-source desktop AI assistant. Its pages list this strength: Supports a wide range of cloud and local models including OpenAI, Google, Anthropic, xAI, DeepSeek, and Ollama. They also state this limitation: API-based models require an internet connection and a provider API key.
β The AI Tool Bible editorial team
Pros
- β Supports a wide range of cloud and local models including OpenAI, Google, Anthropic, xAI, DeepSeek, and Ollama.
- β The software is free and open source.
- β Includes integrated LlamaIndex RAG for files and webpages.
- β Features built-in Python and OS tools with sandbox and Docker support.
- β Available as a desktop application for Windows, Linux, and macOS.
Cons
- β οΈ API-based models require an internet connection and a provider API key.
Facts
- Platforms
- windows, linux, macos, web (not source-checked)
- Pricing
- free - Free and open-source (MIT); bring your own provider API keys (not source-checked)
- Open source
- Yes (license: NOASSERTION)
- GitHub stars
- 2,023 (checked Nov 13, 2026) (not source-checked)
- Last GitHub push
- Oct 6, 2026 (not source-checked)
- First commit
- Apr 2023 (not source-checked)
- Company
- null (not source-checked)
- Model used
- Multi-model (GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Ollama) (not source-checked)
- Package downloads (30 d)
- 4,311 Β· pypi pygpt-net Β· checked Nov 19, 2026 (not source-checked)
- Earliest Wayback capture
- 2023-07-23 (earliest archived capture of the site, not a launch date)
Where this tool sits
Use cases
Frequently asked
- How much does PyGPT cost?
- PyGPT is free and open-source under the MIT License. You must bring your own provider API keys for cloud models, though local Ollama models can run without external API keys.
- Which AI models does PyGPT support?
- It supports models from OpenAI, Google, Anthropic, xAI, DeepSeek, Perplexity, HuggingFace, OpenAI-compatible APIs, and local Ollama models. This includes GPT-5, Claude, Gemini, Grok, DeepSeek, and Mistral.
- Can PyGPT process local files and documents?
- Yes, PyGPT integrates LlamaIndex RAG to process files, webpages, external sources, media, and conversation history. This allows the assistant to use your local data for context-aware responses.
- What operating systems are supported?
- PyGPT runs on Windows, Linux, and macOS. It is written in Python and functions as a desktop application, providing a local interface for interacting with various AI models.
- Does PyGPT include autonomous agent capabilities?
- Yes, it includes work modes for Agents, Autonomous modes, and Computer use. It also features built-in Python and OS tools, including IPython, system commands, a sandbox, and Docker support.