I have recently released the beta version of Tensor, a service that turns your local ZIM files into vector databases that provides semantic understanding of the contents of your ZIM archive to OpenAI-compatible AI models.

tensor-serve

Combining keyword search and semantic search, Tensor helps produce more accurate and adequate responses for the data you have included in a vector database. Tensor offers a full workflow that is built to enhance the local AI experience, however there is also optional features (offline-first by default) features such as web search for time-sensitive information, and configuring API keys to use (currently working on MCP Server).

This project is still under development, and in the process of optimization. tensor-serve is an open-source project, and everyone is is welcome to use and contribute to this project.

tensor-serve on Github Check out the Github Repository

tensor-serve on PyPi Check out the PyPi Package