How TOLL·402 discovered 25,783 x402 services
The primary CDP Bazaar service crawl, secondary health crawl, normalization, deduplication and labeling rules behind the TOLL·402 datasets.
DIRECT ANSWER
TOLL·402 read Agent Tools only for its 13,637-service comparison threshold, then independently crawled the official Coinbase CDP Bazaar catalog. After normalizing HTTP method and resource URL and removing two shifting-page duplicates, it found 25,783 unique x402 catalog services. A separate 402 Index sweep supplies 74,889 unreviewed route-health records.
Key takeaways
- The target and corpus sources are deliberately different to avoid circular confirmation.
- Every record is labeled discovered/unreviewed and stays outside the curated seed.
- The crawl is reproducible with npm run discovery:crawl and auditable with npm run validate:discovery.
Source and threshold separation
At crawl time, Agent Tools reported 13,637 indexed x402 services. TOLL·402 used that number only as a pass/fail threshold. The independent primary source was Coinbase CDP Bazaar, whose public catalog reported 25,786 service resources at the start of the sweep. The crawl fetched 25,785 rows and retained 25,783 unique normalized method-and-URL entries, clearing the threshold without ingesting Agent Tools records.
Normalization and two corpora
The crawler accepts only HTTP and HTTPS URLs, lowercases hosts, removes fragments, trims non-root trailing slashes and sorts query parameters. It combines the advertised HTTP method with the normalized URL. The primary downloadable corpus contains 25,783 CDP Bazaar services. A separate secondary corpus contains 74,889 402 Index route records with health fields; the two counts are never added together.
- Every record preserves its primary source and crawl time.
- Every bulk record is labeled discovered/unreviewed and remains outside the curated listing seed.
- Descriptions and metadata are treated as untrusted discovery text, never verification.
Known limitations
Coinbase documents each Bazaar catalog entry as an x402-enabled service/resource, so the comparison uses that service-endpoint grain. It is not a count of distinct providers, businesses or domains. Catalog pagination is newest-first and changed during the sweep, which produced two duplicate rows; normalization removed them. Neither corpus proves ownership, output quality, current price, absence of prompt injection or successful settlement.
Related directory entries
Sources and methodology
TOLL·402 distinguishes public claims, registry discovery, unpaid quote checks and settled paid-call verification. Sources below support the visible claims; presence in a registry is not treated as verification.
- TOLL·402 primary service corpus — The 25,783-record normalized CDP Bazaar service dataset.
- Coinbase CDP Bazaar documentation — Primary service catalog definition, public API and pagination rules.
- Agent Tools live statistics — Comparison threshold only; not a crawl source.
- 402 Index health endpoint — Independent secondary route totals and health status.