Use BotsRated From Your Agent
The preflight check before you pay is free and unlimited (normal per-IP rate limits apply), with no key and no account. BotsRated checks a paid agent service before your bot pays it. It covers 4,514 x402 seller hosts on Base and remote MCP servers from the Official MCP Registry (33,122 MCP servers checked, 33,339 listed). Every test result comes with a signed receipt.
1. Try It (Free, No Key)
curl "https://botsrated.com/v1/preflight?url=blockrun.ai"
Read latest_probe (status, valid 402, latency, observed price and payTo), freshness.stale and receipt_url. A 404 with host_not_covered means we have no data. That says nothing bad about the seller.
2. Find A Service For A Need (Free)
curl "https://botsrated.com/v1/find?q=weather"
Returns up to 3 covered x402 sellers and up to 3 remote MCP servers, each with a one-line factual reason and a page link. Each call uses 1 of the 25 free daily checks (shared with /v1/check). Also available as the free MCP tool find. The full ranking with receipts is Best Pick, /v1/best ($0.01).
3. Pick How Your Agent Calls It
| Stack | Install | Guide |
|---|---|---|
| Any MCP client (Claude, Cursor, Strands, ADK, ...) | remote MCP server https://botsrated.com/mcp (streamable HTTP, no auth) | Strands example. Free tools: preflight, find, coverage, sample. |
TypeScript, @x402/fetch or @x402/mcp | npm install @botsrated/x402-preflight | pre-payment guard |
| Vercel AI SDK | npm install @botsrated/ai-sdk ai zod | /docs/ai-sdk |
| Python x402 client | pip install botsrated | PaymentGuard |
| LangChain | pip install langchain-botsrated | tools |
| LlamaIndex | pip install llama-index-tools-botsrated | tool spec |
| Plain HTTP / OpenAPI | no install | openapi.json, llms.txt |
4. Guard Every Payment (Free)
The guards ask the free preflight before your x402 client signs. If the payTo in the 402 was never seen for that host, the guard warns by default; with the strict policy it returns block, so your client does not sign. Unknown hosts are allowed by default. Facts only, no score: you pick a named preset (policies.warn, the default, or policies.strict) and can start with dryRun: true to log decisions without changing anything (carefulBuyer and dry-run-only aliases are shipping next). Live as @botsrated/x402-preflight / botsrated 0.2.0 (createPaymentGuard).
5. Paid Lookups (Opt-In)
/v1/check $0.001, /v1/report $0.002, Best Pick /v1/best $0.01 per call, in USDC on Base via x402. The first 25 checks and 3 reports per client per day are free. The libraries only pay when you pass a payment-enabled fetch and a budget, and they stop at that budget. Free samples of each paid answer: check, report, Best Pick. Plans and alerts: pricing. The advertised example host (blockrun.ai) always shows the 402 so x402 scanners see the payment terms; its free example is at /v1/sample/check.
6. Verify What We Say
Receipts are signed with ed25519 (keys at /.well-known/botsrated-key). The open receipt spec and a stand-alone verifier are at github.com/botsrated/botsrated/tree/main/spec: node verify.mjs <receipt URL>. How we measure: methodology. Wrong data about your service: disputes.
Free Data And Pages
- /v1/pulse: live totals (cached 60 s), never names a seller.
- /v1/rates: measured rates with sample size and source.
- /sellers: browse and filter every covered service and MCP server; one page per host at
/s/<host>. - /what-we-check: what we check for, with the rates we measured.
- /leaderboard: 300 x402 services, ordered by our own recent checks (not a rating).