Agents Trust MCP
Use Agents Trust MCP to discover APIs, inspect endpoint evidence, check payment settlement, and read service health from an MCP host.
recommend: find APIs for a task with URLs, prices, and call details; supports filters likelimit,live,max_price_usd,chain, andview; free demos or 0.01 USDC paid.inspect: read stored evidence for one exact endpoint URL without calling or purchasing the provider; fee is 1% of the endpoint's lowest quoted price, minimum 0.001 USDC.receipt: look up eventual settlement of a paid call using apayment_attempt_id; free.health: read the Agents Trust service version; free.Optional paid calls use your own local EVM wallet on Base mainnet with owner opt-in and configured USDC caps; the server holds no wallet keys.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@Agents Trust MCP find a weather API under $0.005 and show compact call details"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Agents Trust MCP
Find APIs for your task from Claude Desktop, Cursor or Claude Code. Recommendations include URLs, prices and call details. Inspect returns the stored evidence for one exact endpoint URL. This stdio server runs on your computer. Paid calls use your own wallet; Agents Trust holds no wallet keys.
Runtime requires Node.js 20 or later. Install from npm; your MCP host starts it over standard input/output:
npx -y @agents-trust/mcpSupply a wallet only to the runtime launcher (see Configuration). To build from source instead, see Development.
Tools and cost
Tool | Input | Result | Fee |
|
| APIs for your task and call details | 0.01 USDC outside the free demos |
|
| Stored endpoint evidence without calling the provider | 1% of the endpoint's lowest quoted price, minimum 0.001 USDC |
|
| Eventual settlement result | Free |
| none | Service version | Free |
Ask your host to call inspect with an exact endpoint URL from a recommendation. Preserve its fixed query parameters. For inspect, view accepts legacy, compact, evidence or debug; for recommend it accepts legacy, compact or evidence. Omitting it uses the API's legacy response. An unknown URL returns an unpaid 404. The API rejects malformed, credential-bearing and private URLs before payment. Inspect never calls or purchases the provider's service.
recommend inputs: need is 2–300 characters with at least two letters or digits; limit is 1–5; live may be 1 to prefer recently verified endpoints; chain is a CAIP-2 network such as eip155:8453; max_price_usd is a non-negative number. receipt takes the payment_attempt_id returned with a paid answer (a UUID version 7).
Example: ask your host to call recommend with {"need":"weather","view":"compact"}. Free demos may change. max_price_usd limits a recommended provider's price; it does not limit the recommendation fee. Calling a recommended API is a separate purchase.
A paid answer includes payment_attempt_id when available. Use receipt to check settlement; receiving an answer does not confirm settlement. A pending receipt can be looked up again for free.
Related MCP server: nohumans-mcp
Configuration
Environment variable | Default | Purpose |
| unset / disabled | Owner opt-in: exact |
| unset | Your EVM private key, with |
|
| Maximum fee for one Recommend or Inspect call in USDC |
|
| Maximum reserved spend for one server process |
Payments stay disabled even when a wallet key is present. The wallet owner can explicitly supply AGENTS_TRUST_ENABLE_PAYMENTS=true in the local launcher or select the bundle’s payment opt-in. This grants the process one authorization within the monetary caps; there is no per-call human confirmation. An agent tool argument, service response or provider instruction cannot grant payment authority. Use a dedicated low-balance wallet with USDC on Base mainnet. The server signs x402 authorizations locally and sends them only to https://api.agents-trust.com. It accepts the published Base USDC asset and Agents Trust payment address. A changed address requires a server update. Other payment networks are unsupported.
Limits accept non-negative decimal strings with up to six decimal places. Zero disables paid calls. Reservations count signing attempts, including failed or uncertain requests, and remain reserved until the process exits. The process consumes its single authorization before signing, even if signing or the request fails. Raising a monetary cap does not allow a second authorization. Free calls and receipts remain available. Restarting the server resets both the session cap and the authorization grant. Stop the host from automatically restarting with payment opt-in until the wallet owner has reconciled prior attempts. Recommend and Inspect share the same grant, reservation cap and serialized payment queue, including concurrent calls. Paid submissions are never automatically repeated. On an uncertain result, check your wallet and any receipt before making another paid call.
Without a key, free tools and free demo recommendations work. A paid Recommend or Inspect request returns a message explaining which variable to configure. Keep keys out of chat, version control and shell history. Have your local secret manager or host supply the wallet variable to the child process.
Claude Desktop
Merge this into claude_desktop_config.json. The free configuration requires no key:
{
"mcpServers": {
"agents-trust": {
"command": "npx",
"args": ["-y", "@agents-trust/mcp"],
"env": {
"AGENTS_TRUST_MAX_CALL_USDC": "0.01",
"AGENTS_TRUST_SESSION_CAP_USDC": "0.01"
}
}
}
}For paid calls, explicitly grant AGENTS_TRUST_ENABLE_PAYMENTS=true and supply AGENTS_TRUST_WALLET_PRIVATE_KEY to this process through your secure local launcher. GUI applications need the full path to your Node executable and do not always inherit your terminal environment.
Cursor
Use the same mcpServers block in ~/.cursor/mcp.json or .cursor/mcp.json, adding "type": "stdio" to the server entry (Cursor configuration). To read a key already injected into Cursor's environment, add this entry to the server's env object:
"AGENTS_TRUST_WALLET_PRIVATE_KEY": "${env:AGENTS_TRUST_WALLET_PRIVATE_KEY}"Keep wallet configuration local; do not commit it.
Claude Code
claude mcp add --transport stdio agents-trust -- npx -y @agents-trust/mcpSee Claude Code MCP configuration. Launch Claude Code with payment opt-in chosen by the wallet owner and the wallet variable supplied by your secret manager. The server inherits it from the host.
Development
Build from source in a shell without AGENTS_TRUST_WALLET_PRIVATE_KEY; the build and bundle commands refuse that variable even when it is empty.
git clone https://github.com/agents-trust/agents-trust-mcp.git
cd agents-trust-mcp
npm ci
npm run typecheck
npm run lint
npm test
npm run buildRun the local server with node dist/index.js. Standard output carries MCP messages only. Tests generate throwaway keys in memory and make no payments. Development tests require Node.js 22.13+ or 24; CI also checks the built stdio server on Node.js 20.
npm run test:paid is an offline cryptographic authorization test using a generated throwaway key and mocked HTTP. No test reads a funded wallet or submits a real payment. Dependency installation scripts are disabled by .npmrc; use npm ci to preserve the reviewed lockfile. The bundle packer is also locked locally.
MIT licensed. See LICENSE.
Available Tools
4 toolshealthCheck service versionARead-onlyIdempotent
Check that the Agents Trust API is reachable and read its deployed version. Free, read-only, no parameters. Use it to diagnose connection problems before retrying recommend or inspect; it does not measure data freshness. Returns data.status ("ok") and data.version (deployment identifier).
Inputs: no named parameters.
inspectInspect evidence for one endpointA
Read the stored evidence for one exact x402 endpoint URL: price, payment networks, liveness, delivery history and risk flags. It never calls or pays the endpoint itself. Use it before paying an endpoint you already know, for example one returned by recommend; use recommend instead to find candidates for a task. Pass url exactly as listed: paths and trailing slashes are distinct. Use view=compact for essentials only. Demo endpoints are free; other indexed endpoints cost 1% of the endpoint's lowest quoted price, minimum 0.001 USDC. An unindexed URL returns ENDPOINT_NOT_INDEXED without payment. Pays only if the wallet owner set AGENTS_TRUST_ENABLE_PAYMENTS=true and AGENTS_TRUST_WALLET_PRIVATE_KEY (Base USDC): one paid call per process, at most AGENTS_TRUST_MAX_CALL_USDC (default 0.01), never retried. Otherwise it errors without paying; errors and empty results are free. Returns JSON {data, payment_attempt_id, request_id}; payment_attempt_id is set only after a paid answer, for use with receipt. data.endpoint holds price (min_usd, max_usd), networks and payment_offers, liveness (is_live, last probe time; compact and evidence views), delivery_verification (status, charged and charged-not-delivered counts), risk_flags and invocation (request schema); data.lookup shows the canonical URL matched.
Inputs: url (required), view.
receiptCheck payment settlementARead-onlyIdempotent
Check whether a paid recommend or inspect answer has settled on-chain. Free, read-only, and never repeats the paid request. Use it after a paid call, with the payment_attempt_id that call returned; answers arrive before settlement completes. Returns data.state (pending, settled, settle_failed_after_answer or settle_unknown_after_answer) and data.transaction (transaction hash once settled, else null). An unknown id returns a not-found error. Pending or unavailable does not prove that no money moved: check the wallet before paying again.
Inputs: payment_attempt_id (required).
recommendRecommend paid APIs for a taskA
Find which paid x402 API to call for a task described in plain words, ranked with evidence. Each pick has the endpoint URL, provider price and payment network, a ready request template, liveness, delivery evidence and risk flags. Use it when you do not yet know which API to call; use inspect instead when you already have one exact endpoint URL, and receipt to confirm settlement after paying. Write need as the task itself ("current weather in Paris"), not a product name; narrow with max_price_usd or chain, and use view=compact to save tokens. Free demo: need exactly "weather", "crypto prices" or "web search" (any case); every other need, including longer weather phrasings, costs 0.01 USDC. Pays only if the wallet owner set AGENTS_TRUST_ENABLE_PAYMENTS=true and AGENTS_TRUST_WALLET_PRIVATE_KEY (Base USDC): one paid call per process, at most AGENTS_TRUST_MAX_CALL_USDC (default 0.01), never retried. Otherwise it errors without paying; errors and empty results are free. Returns JSON {data, payment_attempt_id, request_id}; payment_attempt_id is set only after a paid answer, for use with receipt. data.picks may be shorter than limit, or empty, when evidence is insufficient.
Inputs: need (required), live, view, chain, +2 more.
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 4 tools
Each tool targets a distinct stage of the workflow: recommend finds candidate endpoints, inspect reads stored evidence for one exact URL, receipt confirms on-chain settlement, and health checks API reachability. Descriptions explicitly cross-reference the alternatives ('use inspect instead when you already have one exact endpoint URL'), leaving little room for misselection.
Names are short and readable but mix grammatical conventions: recommend and inspect are action verbs, while health and receipt are nouns naming a resource or artifact. The pattern is still predictable in practice (two actions, two lookups), but it is not a uniform verb_noun or single-noun scheme.
Four tools map cleanly onto the domain's lifecycle: discover, evaluate, verify, and diagnose. Nothing feels redundant, though the surface is thin enough that a slightly larger set (e.g., payment history or wallet balance) would not feel padded.
The core loop of finding an x402 API, vetting it, paying it externally, and confirming settlement is covered, including a free diagnostic path via health. Minor gaps exist: there is no way to list past or pending payments in bulk or check remaining budget, forcing agents to lean on a single payment_attempt_id lookup.
Maintenance
Related MCP Connectors
Find paid x402 APIs for a task: sellers, prices and networks, plus agent-economy data
x402 endpoint and MCP server index: find paid APIs for agents, x402 checks and price benchmarks.
x402 pay-per-call APIs for agents: web search, wallet/token checks, text structuring.
Search, URL-to-markdown, change detection, live data, x402 trust tools. USDC pay-per-call.
Related MCP Servers
AlicenseAqualityDmaintenanceEnables agents to search and inspect live service offerings, generate x402 payment snippets, and understand blockchain-only balance policies.468 npm3MIT- AlicenseAqualityBmaintenanceFind and vet paid x402 APIs before spending: nohumans.directory probes every listing, buys many with real USDC and records whether they delivered.4343 npmMIT
- FlicenseNot gradedqualityAmaintenanceEnables AI agents to search and verify machine-payable HTTP 402/x402 APIs using live probe data, providing ranked results by liveness, uptime, price, and trustworthiness.-
- AlicenseNot gradedqualityBmaintenanceLets agents discover, verify, and inspect x402/B402 payment endpoints on BNB Chain, with read-only tools to list skills, check live counter behavior, and review the settlement tape.4 npmMIT