XERJ

XERJ is a local AI search tool written in Rust that indexes any folder automatically. The idea is that a coding agent can query the index instead of reading files one by one, which cuts the amount of text it has to process. The same index supports search, retrieval-augmented generation, security audits, and agent memory. It is released under the Apache-2.0 license and is still at a release-candidate stage.

It handles code, documents, logs, and PDFs and offers keyword, hybrid, semantic, vector, and graph-style search. It exposes an Elasticsearch-compatible API, so existing clients and tools can connect without changes. Its headline use case is reference coding, where an agent finds relevant open-source repositories, indexes them, and reuses how they solved similar problems.

The project reports token savings from its own studies and field reports, which are vendor claims rather than independent results. It can be installed with a one-prompt command or run manually, with Windows and other platforms noted in the README. It suits developers using AI coding agents and teams that need fast local search over large folders.