Agent Cookbook
Boring crawl instructions for archive entries, entities, corpus records, changes, schemas, and search. This layer exposes public authorship, provenance, licensing, citation, and reuse fields, and excludes private/internal data.
Crawl Flow
- Fetch
/api/manifest.jsonfor route patterns and schema URLs. - Fetch
/corpus/manifest.json, then/corpus.jsonlfor full ingest. - Fetch
/corpus/chunks.jsonlwhen chunked text is useful. - For updates, fetch
/changes.json?since=YYYY-MM-DDand/corpus/delta.jsonl?since=YYYY-MM-DD. - For a single archive entry, fetch its human, JSON, Markdown, and share package URLs from the manifest patterns.
Example Commands
curl https://vietphanx.com/api/manifest.json
curl https://vietphanx.com/corpus.jsonl
curl 'https://vietphanx.com/corpus/delta.jsonl?since=2026-01-01'
curl 'https://vietphanx.com/search.json?q=field%20log'