Back to the chart

xerj-org/xerj

Local AI search engine and MCP server for coding agents

Chart
Not on a chart right now
Stars
19.3k

Open on GitHubHomepage

What it is

A local-first search engine in a single Rust binary that indexes code, docs, logs, JSONL/CSV, and PDFs for BM25 full-text, vector, and hybrid search. It exposes an Elasticsearch-compatible API and MCP server to give coding agents like Claude Code and Cursor direct code search and agent memory.

Who it is for

Developers using AI coding agents such as Claude Code or Cursor who want to cut token usage and give agents accurate local code and document search.

Why it is trending

It gained 14,186 stars today (19,181 total) following the recent v1.0.0-rc.101 release. The release introduces XERJ Query-1, a CPU-only model that turns questions into short programs over stored facts instead of generating RAG answers with an LLM.

Main features

  • Automatic folder indexing with tree-sitter symbol and line-number extraction
  • BM25 full-text, vector, and hybrid search behind an Elasticsearch-compatible API
  • One-command MCP integration for Claude Code and Cursor plus batch search
  • Offline CPU-only XERJ Query-1 model for programmatic queries over stored facts

Summarized on November 17, 2026 from the README and repository data. It can be wrong; check the source before relying on it. Report a wrong summary