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.
Rust agent-memoryai-agentsbm25claude-codeclicode-searchcursordatabaseelasticsearchfull-text-searchhybrid-searchjsonllocal-firstmcpmcp-servermodel-context-protocolragrustsearch-enginevector-search
Set me up for reference-coding with XERJ and then use it automatically for the rest of our work, so I stop burning output tokens re-deriving APIs I could just look up.
- Make sure a local XERJ instance is running and the
xerj-codeskill is available. If XERJ isn't installed, install it from https://xerj.org/get (curl -fsSL https://xerj.org/get | sh) and start it (xerj -d ./.xerj-data --insecure). - Look at this project — its dependencies, the domains its code touches — and
decide which well-known open-source libraries already solve those problems.
Clone them into a reference corpus with
xerj corpus add <name> <git-url>..., grouped by problem domain (async, parsers, storage, …), not by language. - Index that corpus with
xerj corpus index <name>(it runsxerj autoindex). - From now on, for every coding task this session: before you write
non-trivial code against an unfamiliar, external, or internal-but-uncommon
API, first run
xerj code <corpus> "<what you need>"to retrieve the real implementation, read the returned definition, and adapt it — citefile:linewhen you rely on it, and check the licence before copying. If retrieval returns nothing relevant, say so and fall back to normal work. If you've looped twice on the same error, retrieve before trying a third time. - Re-index whenever the references change; refuse a stale index.
Do this proactively for the rest of the session without me asking again.