xerj
XERJ is a Rust tool that indexes code bases, documents, logs, and PDF files locally, allowing AI agents to search for data. Thanks to its automatic indexing ability, it prevents agents from spending tokens by reading files one by one and finds the code directly related. It supports scanning of code bases, keyword searches, and definition-finding operations over the command line.
xerj-org · Search Engines
Common commercial products in this category: Algolia, Elastic Cloud, Azure AI Search
- You automatically add the code, document, and log files to your folder with one command.
- You shoot directly related code by reducing token consumption during file reading of AI agents.
- You search and find a specific function description along with file and line information.
- • NeutralGitHub Issue
reference-coding: symbol index returns whole class/method bodies and does not index declarations as their own symbols —
“A fair, deterministic retrieval test of the reference-coding claim (grep vs XERJ symbol search over Apache Lucene, 12 verifiable questions, precise symbol queries, coverage gaps held out) found XERJ ingests **32–48x more bytes per answered question than grep** and loses recall…”
- ↓ NegativeGitHub Issue
corpus index: finalize-catalog deadlocks all 297 threads on a fresh build; post-kill resume prints "searchable" while th
“## What happened Standing-corpora rebuild round 3 (fresh `--fresh` build of `xerj-search` against the reference server on :9200, rc.83 binary, ~54k files / 1.28 GB). Two distinct failures, in sequence: ### 1. `finalize-catalog` deadlocks — all 297 threads futex-waiting The index…”
My note
Alternative pages
More in: Search Engines
- Meilisearch★ 59,725 · yesterday
- ZincSearch★ 17,927 · 1 mo ago
- Tantivy★ 16,368 · today
- zvec★ 16,164 · yesterday
- OpenSearch★ 13,937 · today
- flexsearch★ 13,815 · 4 mo ago