zylon-ai/private-gpt is an API layer for AI applications that run on top of local models. It is written in Python, and the latest release is v1.0.1. The project does not run models itself. Instead, it connects to an OpenAI-compatible inference server and exposes capabilities such as message handling and file ingestion through an API. The API structure follows the Claude API.
PrivateGPT collects and searches files and artifacts, and offers search with citations (citation) and agentic RAG. Agentic RAG is an approach in which the model composes retrieval and tool calls on its own to produce an answer. Tools, MCP connectors, and database and CSV access are supported as well. The repository description says it also includes skills and text-to-sql. The built-in Workbench UI is for testing and demos; the API is the core product. It was created on May 2, 2023 under the Apache-2.0 license, and has 57,598 GitHub stars.
Still, what can be confirmed is limited to the feature scope and repository metrics. The given record contains no changelog for v1.0.1, so it is not possible to know what this release changed. No performance figures or benchmarks are presented either. On the 10-point scale, only the momentum score is given, at 2.7, while the freshness, technology, and quality scores are empty. The API structure and supported scope can therefore be confirmed, but judgments grounded in actual performance or quality are hard to draw from this record alone.