CONTRIBUTE·ADD YOUR CORPUS

ADD YOUR
CORPUS

Sharing data is like sharing code: it compounds. A corpus you contribute becomes reference memory that any agent can search, so your domain knowledge makes everyone's agents better, not just yours.

Contribute a corpus

Paste this into your coding agent. It explains itself.

Help me contribute a corpus to the XERJ Corpus Hub (hub.xerj.org).
1. If the xerj CLI is not installed, install it: curl -fsSL https://xerj.org/get | sh
2. Build a corpus from a folder of documents I point you to: xerj autoindex <folder>
   (or from public repos: xerj corpus add <name> <git-url> ...). This pins the sources
   (commit SHA, files, bytes) and writes a manifest.
3. Confirm the content is openly licensed and attributable. The hub is Apache-2.0 and
   accepts only openly-licensed, source-cited corpora; record a review block per source
   rather than trusting the licence detector.
4. Fork github.com/xerj-org/xerj, copy the manifest to tools/xerj-code/hub/<name>.json,
   add a one-line row to tools/xerj-code/hub/backlog/backlog-100.json, and open a pull
   request targeting the corpus-hub branch. In the PR, say what the corpus is FOR.

Full guide, templates and the review checklist: CONTRIBUTING.md.