Hatcher ecosystem guide

Public evidence boundary

The Hatcher integration publishes only public ecosystem evidence. It never imports private applicant, investor, contact, or deal-room data. Every item records source URL, observation time, hash, parser/version, and public classification.

Score inputs

The public Hatcher score is a documented 100-point rubric over bounded evidence categories. Missing evidence is missing—not negative evidence—and reduces coverage/confidence. A score includes rubric version, component points, source references, observation times, and freshness.

{
  "rubricVersion": 1,
  "score": 72,
  "coveragePpm": 800000,
  "components": [
    {"code":"PUBLIC_PRODUCT_EVIDENCE","points":24,"maximum":30},
    {"code":"PUBLIC_TECHNICAL_EVIDENCE","points":20,"maximum":30}
  ]
}

Partner data is a source signal, not an ownership, verification, endorsement, or investment claim. Public pages label Hatcher provenance and link to the score methodology. Corrections append a new observation/projection version rather than rewriting evidence.

Synchronization

The scheduled worker uses conditional fetch/cache, body and time limits, deterministic item keys, and at-least-once idempotency. Provider failure serves the last validated projection with stale status where safe. Operators can inspect source lag and replay a bounded failed observation without calling unrelated providers.