State of the x402 Bazaar
The first independent audit of every x402 payment service. 405 endpoints probed for protocol compliance, contract consistency, and schema honesty.
Executive Summary
We independently probed every x402 payment service listed on the Bazaar - 405 unique domains - testing whether each one actually speaks the x402 protocol, returns valid payment data, and delivers what its listing promises. This is the first audit of its kind.
This audit reflects the state of the Bazaar on February 10-11, 2026. Live trust scores are updated continuously and available on the leaderboard.
Methodology
ScoutScore's fidelity scanner sends unauthenticated HTTP requests to each service's listed endpoint URL. For each service, we check:
- 1.Protocol compliance - Does the endpoint return HTTP 402 with valid x402 headers?
- 2.Contract consistency - Do the price, wallet, and network match the Bazaar listing?
- 3.Response structure - Is the response valid JSON? Does it include payment instructions? Does the claimed schema actually exist?
Each service receives a fidelity score (0-100) composed of three layers: protocol compliance (40%), contract consistency (35%), and response structure (25%). Services also receive binary flags (PROTOCOL_COMPLIANT, SCHEMA_PHANTOM, PRICE_MISMATCH, etc.) that identify specific issues.
All 405 x402 domains on the Bazaar were probed between February 10-11, 2026. MCP servers (1,229 GitHub repositories) were not included in this scan - they use a different protocol and require separate methodology.
Finding 1: 57% of Services Don't Work
Of the 405 x402 services listed on the Bazaar, 232 (57%) do not return an HTTP 402 response. They return no payment data, no x402 headers, and no payment instructions. They are listed on the Bazaar but functionally absent.
The compliant 43% return proper 402 responses with valid payment data. These are real, functional x402 services. The non-functional 57% are a mix of abandoned projects, broken deployments, moved endpoints, and services that were never fully implemented.
Finding 2: 57% of Listed Schemas Are Phantoms
228 services (56% of the Bazaar) claim to have output schemas in their listing. When we probe the actual endpoint, 130 of them (57%) deliver nothing - no schema, no documentation, no structured output definition.
We call these "schema phantoms" - services that advertise output schemas they don't have. Only 40% of services that claim schemas actually deliver them when probed.
This matters because agents rely on schemas to validate responses. If an agent expects structured output and receives unstructured data (or nothing at all), it can't verify what it paid for. Schema phantoms create a false sense of predictability in the catalog.
Finding 3: 58% Have Price Mismatches
We compared each service's Bazaar listing (the price, wallet, and network advertised in the catalog) against the actual values returned by the live endpoint.
| Check | Match | Mismatch | Mismatch Rate |
|---|---|---|---|
| Price | 171 | 234 | 58% |
| Wallet | 400 | 5 | 1.2% |
| Fully verified contract | 149 | - | 37% pass all checks |
Price mismatch is the most common contract issue. 58% of services have endpoint prices that differ from what their Bazaar listing advertises. This means an agent checking the catalog price before deciding to pay may be working with incorrect information.
The good news: wallet routing is solid. Only 1.2% of services have wallet address mismatches between listing and endpoint. When an agent sends payment, it almost always goes to the right place. The problem is knowing what you're paying - not where the payment goes.
Finding 4: V2 Adoption Is at 16%
The x402 protocol has two versions. V1 uses JSON body responses. V2 uses the PAYMENT-REQUIRED header with CAIP-2 network identifiers - a more standardized approach.
Of the 173 protocol-compliant services, 145 (84%) use V1 and 28 (16%) have adopted V2. V2 adoption is still early but growing. We will track this weekly as a measure of ecosystem maturation.
Finding 5: Base Dominates, Micro-Pricing Prevails
Network Distribution
Price Distribution
Base mainnet is the dominant network for x402 services, hosting 84% of all listings. 11% are still on the Base Sepolia testnet. Nearly half of all services price below $0.01 per call - the micro-transaction model that x402 was designed for.
Availability: 85% Online Across the Full Directory
Separately from the fidelity audit, we health-checked 1,580 services across our entire directory (x402 and MCP combined). A health check tests whether the endpoint responds to an HTTP request within 10 seconds.
Average response time for healthy services is 1,009ms. Being online is necessary but not sufficient - a service can be "UP" and still fail every fidelity check. That is why we score both availability and fidelity independently.
Ecosystem Health Grade
Average fidelity score: 40.3 / 100
The x402 Bazaar is functional but unreliable. A minority of services work correctly. The majority are either broken, inconsistent, or misleading. The ecosystem needs better quality signals before agents can transact with confidence.
What This Means
For agents and developers: Do not trust Bazaar listings at face value. A service being listed does not mean it works, and a service claiming a schema does not mean it has one. Before your agent makes an x402 payment, check the service's trust score.
For service providers: If your service is protocol-compliant and contract-verified, you are in the top 37% of the ecosystem. That is a competitive advantage worth showing. If your service has issues, the most common fixes are: update your listed price to match your endpoint, and add the output schema your listing claims to have.
For the ecosystem: The x402 payment rails on Base have processed 124.3M transactions totaling $37.2M USDC. The rails work. But 57% of listed services don't function, 58% have price mismatches, and there are only 469 reputation reviews across 22,671 registered agents in ERC-8004. The trust layer does not exist yet. As agent commerce grows, the cost of interacting with broken or dishonest services grows with it.
Check Any Service
Every data point from this audit is available through our free API. No API key required. No rate limits during launch.
About ScoutScore
ScoutScore is the trust layer for agent-to-agent commerce. We independently probe, verify, and score every service in the x402 ecosystem so agents know who to pay before sending a dollar. Our scoring model evaluates contract clarity, availability, response fidelity, and identity signals - the behavioral equivalent of a credit score for AI agents.
This report was generated from ScoutScore's proprietary fidelity scanning infrastructure. All data is available through our public API at scoutscore.ai.