x402 Endpoint Trust Index: August 2026
The August snapshot of x402 resource evidence by network and provider, with exact-route quote validity, freshness and paid-call status, compared against July.
DIRECT ANSWER
The August 2026 index covers 89,057 canonical resources, of which 77,809 were eligible for an exact-route quote check and 19,259 returned a valid x402 payment requirement. No resource in this edition carries a paid-call verification. Provider concentration eased: the largest provider now holds 46.5% of eligible resources, down from 51.8% in July.
Key takeaways
- Valid quote coverage rose to 19,259 confirmed routes, up from 13,928 in July.
- Concentration fell across every band, so the corpus depends less on one provider than it did a month ago.
- Paid-call verification remains at zero for the whole index; a valid quote is not a settled payment.
What this edition measures
Each resource is one canonical uppercase HTTP method plus a normalized URL. A valid quote means that exact route returned HTTP 402 requirements which passed protocol parsing and validation on the recorded date. It is not a paid call, and it does not assert that the paid response would be useful.
The August snapshot was built from source data generated on 5 August 2026 and reported at an evidence age of 0.8 days. Every figure below is scoped to that snapshot. Earlier editions stay published unchanged so a reader can compare months rather than watch a single page silently rewrite itself.
Coverage and quote outcomes
The index classified a quote outcome for all 89,057 canonical resources. Of those, 77,809 were eligible for an automated exact-route check, and 75,604 checks were actually attempted. The gap between eligible and attempted is made up of routes skipped because their origin was unreachable at check time.
Confirmed valid quotes reached 19,259. The remaining attempted routes divide into responses that were not 402 at all, requirements that failed validation, and routes that could not be called safely. A non-402 result is the largest single outcome and usually means the route no longer presents a payment wall at the advertised method and path.
| Outcome | August 2026 | Meaning |
|---|---|---|
| Confirmed | 19,259 | Exact route returned valid requirements |
| Invalid 402 | 1,181 | Payment response failed protocol validation |
| Unparsed 402 | 170 | 402 returned but requirement could not be decoded |
| Non-402 | 54,954 | Route did not present an x402 offer |
| Not safely testable | 11,248 | Mutating method or unresolved template |
| Origin unreachable | 2,205 | Skipped before the route was called |
| Unreachable | 40 | Network or DNS failure at check time |
How August compares with July
The July edition covered 78,315 resources with 13,928 confirmed quotes. August covers 89,057 with 19,259 confirmed. Both the denominator and the numerator grew, and confirmed coverage as a share of eligible resources improved rather than being diluted by the larger corpus.
Readers comparing against the July article should note that the July snapshot was itself rebuilt to an as-of date of 31 July, which moved its resource total slightly. The published July edition now reflects that rebuild. Comparing narrative figures across editions is less reliable than comparing the snapshots themselves, which is why both remain downloadable.
| Measure | July 2026 | August 2026 |
|---|---|---|
| Canonical resources | 78,315 | 89,057 |
| Eligible for quote check | 69,886 | 77,809 |
| Confirmed valid quotes | 13,928 | 19,259 |
| Paid-call verified | 0 | 0 |
| Top provider share of eligible | 51.8% | 46.5% |
Concentration eased but remains high
One provider still accounts for 46.5% of eligible resources, three account for 64.8% and ten account for 72.3%. Every one of those bands fell from July, where the same figures were 51.8%, 72.1% and 80.2%.
The direction is healthy and the level is not. Any average computed per route across this corpus still describes the largest sellers more than it describes a typical one. That is the reason the rate card collapses templated route families into distinct offers and lets each provider contribute its own median exactly once.
Freshness
Evidence age uses the newest recorded source observation for each resource. In this edition 89,053 of 89,057 resources carry evidence recorded within seven days, and all 89,057 within thirty days. That is a property of the refresh having re-verified the corpus rather than carrying old outcomes forward.
Freshness is not the same as correctness. A recently recorded non-402 outcome is fresh evidence that a route did not present a payment wall. Readers should treat the date as telling them when the observation was made, not how confident to be in the service behind it.
Paid-call status across the index
Paid-call verified stands at zero, and reviewed paid output stands at zero, for every resource in this edition. Nothing in the index should be read as evidence that a payment settled or that a paid response matched its documentation.
This is a deliberate reporting choice rather than a gap waiting to be quietly filled. Paid verification is narrow, expensive and route-specific; publishing a zero is more honest than publishing a number assembled from provider self-reports.
- Provider reports live: no payment, no output check
- Quote checked: no payment, protocol shape validated
- Paid-call verified: payment authorized and output compared with documentation
- Index status this month: quote evidence only
Networks and missing identifiers
The August snapshot resolves 38 distinct networks, one more than July. Network is taken from the advertised payment requirement rather than inferred, so a resource whose requirement omits or malforms the identifier is recorded as missing rather than guessed into a bucket.
329 resources carry no resolvable network identifier this month, up from 287 in July. The absolute rise tracks corpus growth, and the category is worth watching because a missing network identifier is a concrete interoperability defect a provider can fix.
How to use this edition
For a buyer, the useful column is the quote outcome on the specific route being considered, not the index total. A corpus-level confirmation rate says nothing about whether one endpoint will answer.
For a provider, the useful signal is which outcome your own routes landed in. Invalid and unparsed 402 responses are protocol defects with a clear fix. Not safely testable is not a failure and needs no action beyond understanding why an automated crawler declined to call the route.
July 2026 Endpoint Trust Index data
AS OF 2026-08-05This is a transparent evidence index, not a single composite score. Network counts overlap when one resource advertises multiple networks. A valid quote is not a settled paid call, and zero paid-call records does not mean zero services can settle.
The largest provider group accounts for 46.5% of quote-testable routes; the ten largest account for 72.3%. Read the network and provider rates with that concentration in mind.
- Canonical resources
- 89,057
- Provider groups
- 3,254
- Quote-testable routes
- 77,809
- Valid / quote-testable
- 19,259 · 24.8%
- Success among attempts
- 25.5%
- Evidence within 7 days
- 100%
- Settled paid calls
- 0 · 0%
- Paid outputs reviewed
- 0 · 0%
Evidence by normalized network
| Dimension | Resources | Quote-testable | Valid / testable | Evidence ≤7d | Paid call |
|---|---|---|---|---|---|
| Base | 83,917 | 73,081 | 21% | 100% | 0% |
| Solana | 11,988 | 9,824 | 95.4% | 100% | 0% |
| Polygon | 2,586 | 1,822 | 97.9% | 100% | 0% |
| Arbitrum | 2,290 | 1,644 | 99.1% | 100% | 0% |
| Algorand | 1,381 | 1,016 | 99.3% | 100% | 0% |
| eip155:143 | 1,362 | 983 | 99.3% | 100% | 0% |
| eip155:480 | 1,200 | 1,158 | 99.4% | 100% | 0% |
| eip155:196 | 1,029 | 957 | 99.3% | 100% | 0% |
| eip155:56 | 950 | 902 | 99.3% | 100% | 0% |
| xrpl:0 | 944 | 921 | 92.3% | 100% | 0% |
Evidence for the 15 largest provider groups
| Dimension | Resources | Quote-testable | Valid / testable | Evidence ≤7d | Paid call |
|---|---|---|---|---|---|
| orbisapi | 36,884 | 36,169 | 0% | 100% | 0% |
| lowpaymentfee | 10,662 | 10,662 | 0% | 100% | 0% |
| deepnets | 3,581 | 3,570 | 99.9% | 100% | 0% |
| Agent402 | 1,900 | 1,496 | 8.2% | 100% | 0% |
| tcgapi | 1,231 | 1,231 | 100% | 100% | 0% |
| clonecho.builda.company | 993 | 993 | 2.4% | 100% | 0% |
| x402 | 830 | 415 | 100% | 100% | 0% |
| GoCreative AI | 751 | 719 | 98.9% | 100% | 0% |
| library | 632 | 316 | 100% | 100% | 0% |
| proxy | 557 | 24 | 45.8% | 100% | 0% |
| x402node | 540 | 540 | 98% | 100% | 0% |
| 2s.io | 490 | 490 | 98.8% | 100% | 0% |
| llm402.ai | 475 | 0 | 0% | 100% | 0% |
| stratalize | 387 | 199 | 100% | 100% | 0% |
| agentutility | 379 | 0 | 0% | 100% | 0% |
EXACT-ROUTE OUTCOMES
- non 402
- 54,954 · 61.7%
- confirmed
- 19,259 · 21.6%
- not safely testable
- 11,248 · 12.6%
- skipped origin unreachable
- 2,205 · 2.5%
- http 402 invalid
- 1,181 · 1.3%
- http 402 unparsed
- 170 · 0.2%
- unreachable
- 40 · 0%
EVIDENCE AGE
- Within 7 days
- 89,053 · 100%
- 8 to 30 days
- 4 · 0%
- Over 30 days
- 0 · 0%
- No recorded date
- 0 · 0%
Source dataset generated 2026-08-05T04:24:21.954Z; it was 0.8 days old at this report's cutoff. SHA-256 7d8b1e2cf375… Download the monthly snapshot history →
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 endpoint trust index — The published monthly snapshot this edition reports.
- TOLL·402 discovery corpus — Resource-level records the index aggregates.
- x402 protocol documentation — Network and asset identifiers used to validate payment requirements.
- x402 rate card — Price benchmarks built from the same August refresh.