The Coinbase x402 Bazaar is the de-facto registry for x402-payable services. Nobody publishes an independent, dated snapshot of what is actually in it. This page is that snapshot: full-pagination capture of the public discovery endpoint at 2026-10-05T17:13Z, derived into one row per listing, with aggregate metrics (host concentration, network split, scheme usage, price distribution, listing recency, field completeness) and a 100-URL seeded liveness probe. The purpose is to give builders, funders, and researchers a citable baseline — one dated row they can anchor claims to about the size and shape of the x402 economy. The raw CDP payload is 215 MB and not published; the dataset.csv (one row per listing, 21 columns) and the summary.json aggregates are. CC BY 4.0.
- One host holds 39% of the entire index.
market.datapackvibe.comaccounts for 13,481 of the 34,587 listings. The top 10 hosts together hold 57.91%. Herfindahl–Hirschman on listings per host is 0.1571 — concentrated but not single-vendor. - The median listed price is $0.01. 83.55% of listings price at or below one cent per call; 96.55% at or below ten cents. The p99 price is $0.945. Microtransactions, not SaaS seat pricing, are the modal unit.
- Only 32 of 34,587 listings are MCP-type. The Bazaar is overwhelmingly HTTP, not MCP-first — 99.907% of the index is a classical URL you call directly. If you came here for MCP distribution, the Bazaar is the wrong lever.
Our own disclosure: zero of the 34,587 listings resolve to onchainagentintel.io. We publish /.well-known/x402.json on our own domain and have not registered with the Bazaar as of the fetch window. This is stated here plainly — a self-published dataset should not hide a disclosure that reflects on the publisher.
Headline table
| Metric | Value | Notes |
|---|---|---|
| Total listings | 34,587 | From pagination.total on page 1; 173 pages at 200/page; truncated=false. |
| Distinct hosts | 2,122 | Unique URL netloc across all listings. |
| Top-1 host share | 38.98% | market.datapackvibe.com — 13,481 listings. |
| Top-10 host share | 57.91% | Of the 34,587 total. |
| HHI on hosts | 0.1571 | Concentrated but not single-vendor. |
| Resource type | http: 34,555 (99.907%) · mcp: 32 (0.093%) | MCP-type listings are a rounding error. |
| Mainnet-only listings | 32,216 (93.14%) | Mainnet-accepted: 99.47%. Testnet-accepted: 6.86%. |
| Dominant payment network | Base mainnet (eip155:8453) | 35,534 of 59,803 accepts[] entries; Solana 7,359; Polygon 3,313; Arbitrum 2,527. |
| Median listed price | $0.01 USD | P90 $0.03, P99 $0.945, max $1,000.00. |
| Listings ≤ $0.01 | 28,817 of 34,490 priced (83.55%) | 97 listings have no inferable USD price (not USDC/USDT/stable, no decimals). |
| Fresh in last 7 days | 26,053 of 34,587 (75.33%) | lastUpdated within 7 days of fetch window. |
| Description populated | 99.82% | Service name populated: 44.4%. |
| Liveness probe (n=100) | 40% HTTP 402 · 47% HTTP 405 · 1% HTTP 200 · 7% net error · 5% other 4xx | GET-only probe, 5s timeout, seed 20261005. 405 = POST-only resource. |
Top 20 hosts
Every value below is verbatim from summary.json (host_summary.top20_hosts). The top ten alone are 57.91% of the entire index.
| # | Host | Listings | Share |
|---|---|---|---|
| 1 | market.datapackvibe.com | 13,481 | 38.98% |
| 2 | agentshelf.syntexa.ch | 1,182 | 3.42% |
| 3 | api.vextorium.com | 1,014 | 2.93% |
| 4 | intel.rallylive.ca | 892 | 2.58% |
| 5 | api.magentlab.com | 844 | 2.44% |
| 6 | dns.intel.rallylive.ca | 698 | 2.02% |
| 7 | agent402.tools | 604 | 1.75% |
| 8 | pay.edge-agents.ai | 586 | 1.69% |
| 9 | api.x402node.dev | 366 | 1.06% |
| 10 | twin.unykorn.org | 361 | 1.04% |
| 11 | email.intel.rallylive.ca | 282 | 0.82% |
| 12 | x402-factory.com | 189 | 0.55% |
| 13 | agents.daedalusdevelopmentgroup.com | 183 | 0.53% |
| 14 | api.myceliasignal.com | 176 | 0.51% |
| 15 | 402.com.tr | 173 | 0.50% |
| 16 | saylorinnovations.com | 172 | 0.50% |
| 17 | api.xynaptic.io | 148 | 0.43% |
| 18 | api.strale.io | 145 | 0.42% |
| 19 | api.agentstools.dev | 141 | 0.41% |
| 20 | minia2a.uk | 140 | 0.40% |
Downloads
| File | Type | Size | SHA-256 |
|---|---|---|---|
| dataset.csv | text/csv | 8.5 MB | fdafde106ff214063dfccefce07e6132151c66c203510ac77e6ccd8022e65aea |
| summary.json | application/json | 7,031 B | 2c2d034992c82172ef840cf6b027a7bd975b5b10c6dc398e81d552885d52e069 |
| dataset.jsonld | application/ld+json | 6,668 B | 1c468d67cd53c22039909b18d4a73239f6b4cd7bebf1ae0e0581d496a2e5a2d4 |
| liveness_probe.json | application/json | 18,795 B | 15dc580d3f3b4472d78bd44c458b02b98249a3a7f9626f97f429cba5a2bb0345 |
| snapshot_meta.json | application/json | 538 B | e9d3d5fdbd5aa89e32471b6c51eea1a929e38da1bb5f9b165635e75ce239d06f |
| PROVENANCE.json | application/json | 3,264 B | 7436a347f4c167ee49d9194490d778e08152adcb7452c075d9f82b48639115a7 |
| SHA256SUMS | text/plain | — | — |
Weekly delta
This dataset is a dated 2026-10-05 baseline. We also publish a recurring weekly recomputation against the same source endpoint — the full set of weekly entries lives under weekly/MANIFEST.json (one entry per snapshot_date, newest last). Numbers below are verbatim from weekly/2026-10-14/summary.json and weekly/2026-10-14/delta.json — the latest entry in that manifest; a self-published dataset should not hide its own disclosure, and the gap between “listing” and “liveness” is still the verification work.
| Metric | Latest week (2026-10-14) | vs. 2026-10 baseline |
|---|---|---|
| Total listings | 34,412 | −175 (baseline 34,587) |
| Top-1 host share | 28.72% (market.datapackvibe.com — 9,884 listings) | −10.26pp (baseline 38.98%) |
| HHI on hosts | 0.08771 | −0.069403 (baseline 0.1571) |
| Top-10 host share | 46.79% | −11.12pp (baseline 57.91%) |
| Week-on-week churn | 51.7773% (8,950 added, 8,865 removed) | vs. 2026-10-07 (34,648 items) |
The 2026-10-14 snapshot adds a quality_summary block with the Bazaar’s per-resource facilitator fields (l30DaysUniquePayers, l30DaysTotalCalls, lastCalledAt). Caveat verbatim from summary.json: “quality fields (l30DaysUniquePayers, l30DaysTotalCalls, lastCalledAt) are facilitator-reported and NOT filtered for self-purchases or test wallets — same caveat Coinbase’s FAQ gives. Catalog accounting, not endorsement.”
The baseline bytes are frozen (sha256 2c2d034992c82172ef840cf6b027a7bd975b5b10c6dc398e81d552885d52e069, pinned in delta.json.baseline); each weekly file is written to a dated subpath and never mutates in place. Full artefact index: weekly/MANIFEST.json.
Reading
| File | Purpose |
|---|---|
| DATA-DICTIONARY.md | Column-by-column reader's map (21 columns) and the aggregate fields in summary.json. |
| METHOD.md | Fetch window, pagination, price-inference and testnet-classification rules, liveness-probe design, limitations. |
| PROVENANCE.json | Snapshot-time ledger: endpoint, fetch window, raw sha256, inference rules, limitations, explicit disclosure that our own endpoints are not listed. |
FAQ
- What is the Coinbase x402 Bazaar?
- A public discovery index operated by Coinbase CDP that lists services accepting payment via the x402 protocol (HTTP 402 Payment Required with on-chain settlement). The index exposes a public, unauthenticated discovery endpoint at
/platform/v2/x402/discovery/resourcesthat returns one record per listed resource, each with pricing, networks, schemes, and an input/output schema. - How many services are in the Coinbase x402 Bazaar?
- As of 2026-10-05T17:13Z, 34,587 listed resources across 2,122 distinct hosts. One host (
market.datapackvibe.com) holds 13,481 of those listings — 38.98% of the entire index — and the top 10 hosts together hold 57.91%. Only 32 of the 34,587 listings are typedmcp; the rest arehttp. - How do services get listed in the Coinbase x402 Bazaar?
- Listing is initiated by the service provider via the Coinbase CDP x402 facilitator; typical paths include registering the resource through the CDP developer console, declaring it in the x402 facilitator, or submitting it through the x402 submission flow. Publishing a
/.well-known/x402.jsonon your own domain is not sufficient on its own — a service that only publishes/.well-knownis not visible to the Bazaar discovery endpoint until a submission is made. This census found zero listings pointing atonchainagentintel.ioas of the fetch window. - What does an x402 service typically cost?
- Median listed price is $0.01 USD per call; p90 $0.03; p99 $0.945. 83.55% of listings price at or below one cent; 96.55% at or below ten cents. Only a handful exceed $1.00; the maximum observed price is $1,000 (likely a one-shot workload or a mis-flagged batch unit). Prices are quoted as the cheapest USD-denominated accepts[] entry on each listing; see METHOD.md for the inference rules.
- Which networks dominate x402 payment settlement?
- 99.47% of listings accept at least one mainnet, and 93.14% are mainnet-only (no testnet accepts entry). The modal network is Base mainnet (
eip155:8453) with 35,534 of 59,803 accepts[] entries. Solana mainnet is second (7,359), then Polygon mainnet (3,313), Arbitrum One (2,527), Base Sepolia (2,362), and Optimism mainnet (1,178). The index spans 47 distinct CAIP-2 network identifiers in total, including XRPL, Stellar pubnet, Algorand mainnet, and multiple less-common EVM chains. - How many of the listed x402 services actually answer the phone?
- A seeded random sample of 100 URLs (seed
20261005) probed with a single HTTP GET and a 5-second timeout returned: 40% HTTP 402 (the correct x402 payment challenge), 47% HTTP 405 Method Not Allowed (resource is POST-only), 1% HTTP 200, 5% other 4xx (401/404), and 7% network errors (timeout / URL error / connect reset). This is a lower-bound liveness check — it does not confirm any particular listing delivers the capability it claims after payment.
Cross-reference
Narrative companion piece on this snapshot: We indexed all 34,587 x402 Bazaar resources. 39% come from one host, ~1% return 200. — walks through methodology and six findings; every number cited is a named field in summary.json.
If you want the on-chain view of the agent economy (ERC-8004 IdentityRegistry registrations and distinct operators across four chains), see the ERC-8004 four-chain census (2026-09-17). The two datasets measure overlapping but distinct populations: the x402 Bazaar is about who takes payment, the ERC-8004 census is about who has registered an on-chain identity. Together they frame the x402 + ERC-8004 agent economy from both sides.
If you're a service provider: our developer-facing docs at /docs cover how we expose /.well-known/x402.json, how our paid intelligence API uses x402, and the Safe-controlled payment wallet.
If you need a stable columnar bulk-extract of our ERC-8004 datasets (chain-liveness and commerce-backed) with weekly Parquet delivery and a semver schema guarantee, see the dataset-licence schema & changelog (v1.0.0). Free CC BY 4.0 content on this page is unaffected.
Cite as
Onchain Agent Intel (2026). Coinbase x402 Bazaar discovery census (2026-10-05). CC BY 4.0. https://onchainagentintel.io/data/x402-bazaar-census-2026-10
License
This dataset is published under Creative Commons Attribution 4.0 International (CC BY 4.0). Free to reuse, remix, and redistribute — including commercially — with attribution.