Agent & Infrastructure

Audio transcription API for agents: paid per call, no keys, no accounts. Neural speech-recognition transcript of any public audio URL up to 25MB, with segment text. Nothing retained.

Returns current or historical cryptocurrency price data.

Latest-release metadata from the official PyPI JSON API (live): version, license, requires_python compatibility, requires_dist dependency specs (max 100), yanked flag, last upload time and project URLs. Query: ?pkg=requests. No personal author/maintainer data. 1h cache.

Paid single-attempt reasoning puzzle for AI agents at The Latent Lounge — dead-reckon a robot's final grid coordinates from a stream of moves. Every puzzle is freshly generated per request (no fixed test set to game), with automatic answer checking. Correct solves build streaks and rank you on the public agent leaderboard. Standard tier.

Cryptographic verification of the offer-receipt extension of x402. Accepts a signed offer in eip712, jws or eip191 form — either bundled as {format, payload, signature} or as payload plus a separate signature — and an optional signed receipt. Recovers the signer, compares it with the payTo of the offer, compares receipt and offer field by field and names every divergence with the value on each side, checks whether the offer has expired and whether the receipt timestamps are plausible against sta

Returns demand depth for one category: probe volume, distinct sources, buyer-class share, probed price points, and a daily time series. Use it to validate pricing before launching in that category.

A stateful duplicate check for agents that settle x402 payments. Supply any of payment_id, payload_hash and nonce; each is hashed with SHA-256 before storage, so the service keeps fingerprints and timestamps and never the values themselves, never a payload, never an address. The answer states for every supplied identifier whether it was seen before, when it was first and last seen, how often, and whether the first sighting falls inside window_seconds. The recommendation is ablehnen only for a re

Fetches the 402 challenge of a foreign endpoint without paying anything — no payment header, no signature, no settlement — or parses a challenge you already hold. Returns x402_version, every accepts entry with scheme, network, resolved network name and chain id, asset, symbol, decimals and the amount both in raw units and in human form, plus the resource block and any expiry. amount_human is computed with the decimals of that token, resolved from a known-asset table or read from the chain; where decimals cannot be established it stays null and is listed in unavailable_fields rather than assuming six. parse_warnings names concrete gaps such as a missing description or missing inputSchema, a non-CAIP-2 network, a bridged instead of canonical asset, or an amount sent as a JSON number. It does not judge the service, does not pay, does not verify signatures and does not check whether the seller delivers. Answers for a URL are cached for 60 seconds; age_ms states the age of the fetched challenge.

A stateless rule check for agents that pay for x402 endpoints. You supply the proposed payment, your policy and the amounts already spent this hour and this day; this endpoint keeps no ledger of its own and returns your counters unchanged. Every configured rule — max_per_call, max_per_hour, max_per_day, allowed_chains, allowed_assets, allowed_recipients, blocked_recipients, require_schema — is reported individually as erfuellt, verletzt, nicht_pruefbar or nicht_gesetzt, together with the limit a

Validates a 402 challenge — fetched from a URL without paying, or supplied directly — against PaymentRequiredV1Schema and PaymentRequiredV2Schema of the installed @x402/core reference implementation. Reports the detected version and how it was detected, every error with field, expected, found and spec_ref naming package, version, schema and field path, whether the network is valid CAIP-2, whether the asset is a known and canonical token address or a bridged variant, whether the declared inputSch

Takes the steps of one agent run and returns a scored report. It counts identical tool calls (same tool, same arguments) as loops, finds the longest run of consecutive failures, marks calls whose result was empty or null as wasted, and attributes cost per step and per tool so the expensive part is visible. Cost is taken from your figures when supplied, otherwise computed from token counts against the model price table.

Wertet robots.txt fuer 23 benannte KI-Agenten aus, mit korrekter Gruppenauswahl und der Regel, dass die laengste passende Anweisung gewinnt. Prueft zusaetzlich X-Robots-Tag, Meta-Angaben, noai, Bezahlschranken-Hinweise, Lizenzverweise, Nutzungsbedingungen und den TDM-Vorbehalt unter /.well-known/tdmrep.json.

Fetches kind 3 contact lists over up to twelve relays. Because kind 3 is replaceable, several relays regularly hold different versions; only the one with the highest created_at is used and the answer states which relays carried that version and how many older versions were seen, so a stale relay never quietly overwrites a current list. At depth 1 the answer is the account and its outgoing follows.

The x402 market by seller over the last 30 days: estimated revenue, paid calls, paying wallets, calls per payer, median price, networks and each seller's top resource - from Coinbase's public x402 discovery index, read hourly. Rank by revenue, calls or payers; filter by keyword and network; a payer floor (default 3) drops lone testers. Send {} for the top 20 by revenue. Pay-per-call via x402.

Fetches https://domain/.well-known/nostr.json with the name query parameter and reads the mapping the domain publishes. The answer states the resolved public key in hex and as npub, the relay hints the document lists for that key, and the HTTP status and final URL after redirects, so a redirect to a login page is distinguishable from a genuine answer. Given an expected public key, the comparison is made and the result is reported as an explicit match or mismatch rather than left to the caller.

Selects objective framing, complication, and reward language from declared templates using a caller seed. It returns a bounded quest outline, not model-authored narrative, encounter balancing, unique lore, scripting, or playable content.

Divides observed CPU, memory-time, network, tool-call and wall-time usage by caller-supplied budgets, reports exceeded dimensions and ranks observations by the maximum ratio. It does not collect telemetry, infer intent, identify inefficient algorithms or terminate an agent.

Searches the public x402 discovery catalogue for services able to replace a broken dependency. Give the failed URL and, better still, a short description of what it did and the fields you need back; matches are scored on overlap between your task words and the candidate name, description and path, then ordered by score and price. Every hit carries the reason it matched, its price in USDC, its network and when the catalogue last saw it.

Performs the MCP handshake against a streamable-HTTP endpoint and then reads the catalogue: tools with their input schemas, resources with their URIs and mime types, and prompts with their arguments. Servers answering as an event stream are handled as well as plain JSON, and a session id is carried across calls. Only listing methods are sent — no tool of the inspected server is ever executed, nothing is written, and no resource contents are read.

Validates an Agent Card document against the A2A convention and returns findings by severity: error (the card is unusable as published), warning (works, but a consuming agent will guess), info (recommended field absent). Checked: the required fields name, description, url, version, capabilities and skills; that url is absolute and https; that every skill carries id, name, description and tags and that skill ids are unique; that declared securitySchemes use a known type and that every entry in…

Gives an agent without a reachable server a place to receive callbacks. create returns a public URL that accepts any POST, GET or PUT from a third party; poll returns the events that arrived since a cursor you carry forward, with headers, parsed body and arrival time. The inbox expires after at most seven days and holds at most 500 events, oldest dropped first. Reading requires the secret returned at creation.

Detect when an Apify Actor, API, scraper or data feed breaks, goes stale or silently changes shape. One paid call returns a verdict with the evidence behind it, so an agent can decide whether a data source is still trustworthy before acting on it. Works with any HTTP endpoint; 35+ pre-wired vendors (OpenAI, Anthropic, Stripe, GitHub) are a head start, not a limit. Schedule it on your own cadence for continuous monitoring of breakage, staleness and drift.

Pay-per-KYA: an agent pays to verify another agent's post-quantum identity + scope credential (Dilithium-2, NIST FIPS 204), with a live FRC-55 on-chain reputation read. ERC-8004-aligned Know Your Agent for autonomous agent-to-agent commerce. x402-native.

Invoice field verification for AI agents: arithmetic consistency of invoice fields, line-item totals, JP registration number check-digit + corporate-number existence observation (official NTA registry), accounting JSON. 0.015/call.