Skip to content

armgit / xerj

Public Imported from GitHub
0

AI search engine and MCP server for bots and agents. One Rust binary indexes code, docs, logs, JSONL/CSV and PDFs for BM25 full-text, vector and hybrid search, code search without grep, agent memory and RAG. Elasticsearch-compatible API; Corpus Hub serves hundreds of reference memory corpora.

688 B
Raw Blame

title: "Where do the XERJ benchmark numbers come from?" evidence:

  • claim: "The board is generated by bench-matrix.mjs" source: "demo/playbooks/SCORECARD.md" expect: [FC-BENCH-SOURCE, FC-BENCH-REPRO, FC-BENCH-P99-FALSE]

Where do the XERJ benchmark numbers come from?

Every number on this page comes from demo/playbooks/BENCHMARK_VS_ES.md and the migration recipe table at docs/recipes/migrate-from-elasticsearch.md:214.

You can reproduce the whole board in four commands on your own machine, and an unexpected LOSE fails CI, so the figures cannot drift.

Under sustained write, Elasticsearch holds p99 at 2-19 ms while XERJ rises to 60-150 ms on a single-node box.