xerj-org/xerj

XERJ is the new way for AI to search data. Its autoindex capability activates agents to know your data without the token waste of grep and sed. One command indexes code, docs, logs and PDFs for search, RAG, security audits and agent memory, using 40x fewer tokens than grep. Elasticsearch compatible, so existing clients just work.

What it solves

XERJ is a local AI search engine designed to prevent coding agents from wasting tokens by reading entire files one by one. It allows agents to retrieve only the exact code passages or documentation they need, significantly reducing input and output token costs while maintaining high solve rates for coding tasks.

How it works

XERJ automatically indexes folders containing code, documentation, logs, PDFs, and other structured data (CSV, JSON, SQLite) without requiring a manual schema. It uses tree-sitter for AST-aware indexing of source code, ensuring that symbols and functions are indexed rather than just flat text. The system provides a REST API compatible with the Elasticsearch API, allowing it to work with existing clients and dashboards. It also supports the Model Context Protocol (MCP) to integrate directly with AI assistants as a tool.

Who it’s for

It is primarily for developers using AI coding agents (like Claude Code or Cursor) who want to reduce token usage and improve agent accuracy on private or niche codebases. It is also suitable for security auditors, log analysts, and developers building RAG applications who need a lightweight, single-binary search backend.

Highlights

  • Reference Coding: Enables agents to index similar open-source projects to find existing solutions before writing new code.
  • Elasticsearch Compatible: Implements the Elasticsearch REST API for seamless integration with existing tooling.
  • Zero-Config Autoindexing: Automatically detects file types and creates indices for any folder.
  • MCP Native: Built-in support for the Model Context Protocol for easy agent integration.
  • Coded-Aware Retrieval: Uses tree-sitter to index symbols and line numbers for precise code retrieval.
  • Hybrid Search: Supports full-text, vector (kNN), and hybrid search in a single query.

Related

  • Project
    elara-labs/code-context-engine

    A local codebase indexing engine that reduces AI coding assistant costs by replacing full-file reads with targeted semantic search and retrieval.

  • Project
    cocoindex-io/cocoindex-code

    An AST-based semantic code search tool that enables AI coding agents and developers to find code by meaning and structure, reducing token usage by up to 70%.

  • Project
    colbymchenry/codegraph

    CodeGraph is a Rust‑based local tool that builds a full‑code knowledge graph for any project (supports dozens of languages) and integrates it with AI coding assistants (Claude Code, Cursor, Copilot, Gemini, etc.). The graph lets agents retrieve exact symbols and call‑paths in a single request, cutting tool calls, file reads, token usage and cost dramatically. Install via a one‑line script, run `codegraph init` per repo, and the CLI auto‑updates the graph as you edit. A browser UI (`codegraph ui`) visualises the graph, and the CLI can uninstall itself cleanly.

  • Project
    jgravelle/jcodemunch-mcp

    A token-efficient MCP server that uses tree-sitter AST parsing to allow AI agents to retrieve precise source code symbols rather than reading entire files.

  • Project
    yvgude/lean-ctx

    A context engineering layer for AI coding agents that compresses inputs and shell outputs to reduce token costs and extend context windows.