Data Quality Gate

A human read this entry. That is the one thing true of everything on the list. Beyond that: the endpoint answered an x402 request when we knocked it, on the date shown. That date is the claim — not a word, a date. None of it is a delivery test. We have not paid for this and graded what came back, so read it as "we checked what is stated above," never as "this is worth the money."

Deterministic post-scrape data cleaner: repairs how data was encoded — residual HTML, mojibake (`Café` → `Café`), invisible characters, non-breaking spaces — never what it says. The 20 rules are published free at `GET /api/clean` before you pay: 7 automatic, 5 opt-in (they change row count, type or schema), 8 report-only — near-duplicates are never merged, failed extractions never deleted. No LLM: identical input yields byte-identical output. `POST /api/clean` $0.04 returns the cleaned data, `/api/clean/audit` $0.12 adds a reversible ledger, `POST /api` $0.01 returns a quality verdict (RELIABLE/USABLE_WITH_CLEANING/UNRELIABLE) with per-check facts. USDC on Base or Solana. ([OpenAPI](https://www.aidatatools.dev/openapi.json)) ([llms.txt](https://www.aidatatools.dev/llms.txt))