# x402-endpoints

> x402 wrappers for official real-world APIs and registries. Each endpoint returns
> data an agent cannot produce on its own: jurisdictional, real-time or official.
> Pay-per-call via x402 (USDC). No API key, no account: payment IS authentication.

Network: eip155:8453 (Base mainnet)
Asset: USDC (0x833589fCD6eDb6E08f4c7C32D4f71b54bdA02913, 6 decimals)
Pay-to: 0x1D1B81247C407521E2A01F3E21514870dcf1620f
Facilitator: https://api.cdp.coinbase.com/platform/v2/x402
Discovery: /.well-known/x402.json — Agent card: /.well-known/agent-card.json

## Endpoints

### GET /gleif/lei
- what: Look up a Legal Entity Identifier (LEI) against the official GLEIF global registry.
  Returns legal name, entity status, jurisdiction, legal form, registered address,
  initial registration date and managing registration authority (LOU). Worldwide coverage.
- input: query param `lei` (required) — 20-character LEI, e.g. `529900T8BM49AURSDO55`
  (pattern `^[A-Z0-9]{20}$`).
- output: JSON. Example: {"lei":"529900T8BM49AURSDO55","legal_name":"Ubisecure Oy",
  "entity_status":"ACTIVE","registration_status":"ISSUED","jurisdiction":"FI",
  "legal_form_code":"DKUW","legal_address":{...},"source":"GLEIF API v1 (gleif.org)",
  "timestamp":"...","cached":false}
- errors: 400 (LEI malformed, ≠ 20 alphanumerics), 404 (LEI unknown),
  502 (GLEIF indisponible), 504 (timeout source).
- price: $0.01 (tier verif)
- source: GLEIF API v1 (gleif.org) — public, free, no key
- health: GET /gleif/health
- llm_usage_prompt: LEI lookup via GLEIF: legal entity name, status, jurisdiction, ultimate & direct parent relationships and registration authority. For regulatory reporting (EMIR/MiFID), KYB and counterparty agents. Input: LEI code or legal name.

### GET /vies/vat
- what: Validate an EU VAT number against the official EU VIES registry across all 27 member
  states. Returns valid/invalid plus the registered company name and address.
- input: query params `country` (required) — 2-letter member state code, e.g. `IE` (use `EL`
  for Greece); `vat` (required) — VAT number without country prefix, e.g. `6388047V`.
- output: JSON. Example: {"country_code":"IE","vat_number":"6388047V","valid":true,
  "name":"GOOGLE IRELAND LIMITED","address":"3RD FLOOR, GORDON HOUSE, BARROW STREET, DUBLIN 4",
  "source":"EU VIES REST API (ec.europa.eu)","timestamp":"...","cached":false}
- errors: 400 (country/vat malformed), 200 valid:false (nonexistent number, legitimate),
  503 (base TVA nationale offline — VIES "unavailable", distinct de invalid), 502/504 (source).
- price: $0.01 (tier verif)
- source: EU VIES REST API (ec.europa.eu) — public, free, no key
- health: GET /vies/health
- llm_usage_prompt: EU VAT number validation via the official VIES registry across 27 member states. Returns valid/invalid plus registered company name and address. Critically distinguishes 'invalid' (number does not exist) from 'unavailable' (national VAT database offline). For invoicing, B2B onboarding, KYB and tax-compliance agents. Input: 2-letter country code (EL for Greece) + VAT number.

### GET /polymarket/odds
- what: Live prediction market odds and implied probabilities from Polymarket. Returns each
  outcome with its probability (0-1), plus volume, liquidity and resolution status.
- input: query param `market` (required) — Polymarket market id or slug, e.g. `2654605`
  or `will-it-rain-tomorrow`.
- output: JSON. Example: {"id":"2654605","question":"...","outcomes":[{"name":"Yes",
  "price":0.06,"probability":0.06},{"name":"No","price":0.94,"probability":0.94}],
  "volume":"99.98","liquidity":"1426.8","source":"Polymarket Gamma API","cached":false}
- errors: 400 (market missing), 404 (id/slug not found), 502/504 (source).
- price: $0.05 (tier verdict)
- source: Polymarket Gamma API (gamma-api.polymarket.com) — public, no key
- health: GET /polymarket/health
- llm_usage_prompt: Live prediction-market odds and implied probabilities from Polymarket. Returns each outcome with its probability (0-1), plus volume, liquidity and resolution status for a market. For event-forecasting, hedging and research agents. Input: Polymarket market id or slug.

### GET /flights/states
- what: Real-time aircraft states worldwide from adsb.fi open data (ADS-B). Returns every aircraft
  in a bounding box with live position, altitude, speed, heading and ground status.
- input: query params `lamin`, `lomin`, `lamax`, `lomax` (all required) — bounding box, e.g.
  lamin=48.0 lomin=2.0 lamax=49.0 lomax=3.0 (must fit within ~250 NM radius from its center).
- output: JSON. Example: {"bbox":{...},"time":1782418394,"count":1,"states":[{"icao24":
  "3c49b1","callsign":"TUI6HA","origin_country":null,"latitude":48.8208,"longitude":2.1897,
  "baro_altitude_m":11582.4,"on_ground":false,"velocity_ms":239.01,"registration":"D-ABMQ",
  "aircraft_type":"B738"}],"source":"adsb.fi open data (opendata.adsb.fi)",
  "attribution":"Live ADS-B data by adsb.fi (https://adsb.fi) — community ODbL"}
- errors: 400 (bbox invalid / trop large > 250 NM), 503 (rate limit adsb.fi 1 req/s), 502/504 (source).
- price: $0.01 (tier verif)
- source: adsb.fi open data (opendata.adsb.fi) — public, no key, ADSBexchange v2 format, ODbL (attribution required)
- health: GET /flights/health
- llm_usage_prompt: Real-time aircraft states worldwide from adsb.fi open data (ADS-B). Returns every aircraft in a bounding box with live position, altitude, velocity, heading and on-ground status. For travel, logistics, airspace-monitoring and tracking agents. Input: bounding box (lamin, lomin, lamax, lomax). Data by adsb.fi (ODbL).

### GET /ev/charging
- what: EV charging stations worldwide from Open Charge Map. Returns nearby stations with
  operator, connector type, power (kW), number of points and status.
- input: query params `latitude` (required), `longitude` (required), `distance` (optional,
  km, default 5), `maxresults` (optional, default 20). E.g. latitude=48.85 longitude=2.35.
- output: JSON. Example: {"query":{...},"count":1,"stations":[{"id":198782,"title":
  "SAEMES | PARKING LAGRANGE","latitude":48.8507,"longitude":2.3487,"connections":[{"type":
  "Type 2 (Socket Only)","power_kw":22.0,"quantity":2,"operational":true}]}],"source":"..."}
- errors: 400 (lat/lng/distance/maxresults out of bounds), 502/504 (source/key).
- price: $0.01 (tier verif)
- source: Open Charge Map (api.openchargemap.io) — key required
- health: GET /ev/health
- llm_usage_prompt: EV charging stations worldwide from Open Charge Map. Returns nearby stations with operator, connector type (Type2/CCS/CHAdeMO), power in kW, number of points and status. For routing, fleet and trip-planning agents. Input: latitude, longitude, radius (km) and max results.

### GET /bodacc/annonces
- what: Search the official French BODACC registry (DILA) for company legal announcements —
  incorporations, modifications, radiations and insolvency proceedings. Coverage France.
- input: query params `q` (required) — company name, SIREN or RCS, e.g. `OVH`; `famille`
  (optional) — creation|modification|radiation|depot|collective; `limit` (optional, default 10).
- output: JSON. Example: {"query":"OVH","total_count":165,"count":1,"annonces":[{"date_parution":
  "2026-06-12","famille":"Créations","commercant":"OVHcloud OCT5","registre":["104274436"],
  "tribunal":"Lille Métropole","url":"..."}],"source":"BODACC — DILA via OpenDataSoft"}
- errors: 400 (q < 2 chars / famille invalid), 502/504 (source).
- price: $0.01 (tier verif)
- source: BODACC — DILA via OpenDataSoft (bodacc-datadila.opendatasoft.com) — public, no key
- health: GET /bodacc/health
- llm_usage_prompt: BODACC company legal announcements monitor (France, official DILA source): incorporations, modifications, radiations and insolvency proceedings for a company. Returns announcement type, family, date, court (tribunal), registre/SIREN and link. For credit-risk, supplier-risk, KYB and AML agents. Input: company name, SIREN or RCS.

### GET /ted/tenders
- what: Search the official EU TED (Tenders Electronic Daily) for public procurement notices
  across the 27+ member states by keywords, buyer country and CPV code.
- input: query params (at least one required) `query` — keywords; `country` — ISO-3 buyer
  country (e.g. FRA); `cpv` — CPV code prefix (e.g. 72000000); `limit` (optional, default 10).
- output: JSON. Example: {"total":31412,"count":1,"tenders":[{"publication_number":"212976-2016",
  "title":"France-Châtenay-Malabry: ...software package","buyer_name":"Ecole Centrale",
  "buyer_country":"FRA","notice_type":"Contract notice","publication_date":"2016-06-22","url":"..."}]}
- errors: 400 (no criteria / country ≠ ISO-3 / non-numeric cpv), 502/504 (source).
- price: $0.05 (tier verdict)
- source: EU TED — Tenders Electronic Daily (api.ted.europa.eu) — public, no key
- health: GET /ted/health
- llm_usage_prompt: EU public-procurement tender search via the official TED (Tenders Electronic Daily). Returns notices: publication number, title, buyer name and country, notice type, publication date and link. Filter by keywords, buyer country (ISO-3) and CPV code. For business-development, bid and market-intelligence agents.

### GET /eurlex/search
- what: Search official EUR-Lex for EU legislation, treaties, case law and legal acts by keyword.
  Returns CELEX number, title and date. Multilingual.
- input: query params `query` (required) — keywords searched in act titles, e.g. `data protection`;
  `language` (optional) — en|fr|de|es|it|nl|pl (default en); `limit` (optional, default 10).
- output: JSON. Example: {"query":"data protection","language":"en","count":1,"documents":
  [{"celex":"32016R0679","title":"Regulation (EU) 2016/679 ...","date":"2016-04-27","url":"..."}],
  "source":"EUR-Lex / CELLAR SPARQL (publications.europa.eu)"}
- errors: 400 (query < 3 chars / langue invalid), 502/504 (source).
- price: $0.01 (tier verif)
- source: EUR-Lex / CELLAR SPARQL (publications.europa.eu) — public, no key
- health: GET /eurlex/health
- llm_usage_prompt: EUR-Lex search for EU legislation, treaties, case law and legal acts via the official EU Publications Office (CELLAR). Full-text title search returns CELEX number, title, date and EUR-Lex link. Multilingual (en, fr, de, es, it, nl, pl). For legal-research, regulatory and compliance agents. Input: keywords.

### GET /sanctions/screen
- what: Screen a name against the official EU consolidated sanctions list (FISMA). Returns
  matches with a similarity score and context — NOT a binary yes/no.
- input: query params `name` (required) — person or entity name, e.g. `Saddam Hussein`; `type`
  (optional) — person|enterprise; `threshold` (optional, default 0.7); `limit` (optional, default 10).
- output: JSON. Example: {"query":"Saddam Hussein","match_count":1,"matches":[{"name":
  "Saddam Hussein Al-Tikriti","score":0.859,"subject_type":"person","eu_reference":"EU.27.28",
  "programme":"IRQ","designation_details":"..."}],"list_size":30242,"source":"EU ... FISMA"}
- errors: 400 (name < 2 chars / type invalid / threshold hors ]0,1]), 502/504 (source).
- price: $0.05 (tier verdict)
- source: EU Consolidated Financial Sanctions List — FISMA (webgate.ec.europa.eu) — public, no key
- health: GET /sanctions/health
- llm_usage_prompt: Name screening against the official EU consolidated financial sanctions list (FISMA). Fuzzy-matches a person or entity name and returns matches with a similarity score plus context: EU reference, subject type, programme and designation details. Not a binary yes/no — returns ranked matches for human review. For KYB, AML and payment-compliance agents. Input: name (+optional type).

### GET /cve/lookup
- what: Security vulnerability lookup (CVE) against the official NVD (NIST) database, by exact
  CVE id or by keyword/product. Returns description, CVSS severity, CWE weaknesses and references.
- input: query params (at least one of) `cve_id` — exact id, e.g. `CVE-2021-44228`; `keyword`
  — product/keyword, e.g. `log4j`; `limit` (optional, default 10, for keyword search).
- output: JSON. Example: {"count":1,"cves":[{"id":"CVE-2021-44228","cvss":{"base_score":10.0,
  "base_severity":"CRITICAL","vector":"CVSS:3.1/..."},"weaknesses":["CWE-502","CWE-917"],
  "status":"Analyzed","references":["..."]}],"source":"NVD — NIST ... v2.0"}
- errors: 400 (neither id nor keyword / cve_id malformed), 404 (CVE not found), 503 (NVD rate-limit), 502/504.
- price: $0.01 (tier verif)
- source: NVD — NIST National Vulnerability Database v2.0 (services.nvd.nist.gov) — public, no key
  (rate-limit ~5 req/30s with no key; a free NVD key raises it to ~50/30s)
- health: GET /cve/health
- llm_usage_prompt: Security vulnerability (CVE) lookup against the official NVD (NIST) database. By exact CVE id or by keyword/product. Returns description, CVSS score and severity, CWE weaknesses, status and references. For devsecops, SBOM and dependency-audit agents. Input: CVE id (e.g. CVE-2021-44228) or keyword (e.g. log4j).

### GET /recalls/search
- what: Product recall lookup (drugs/devices/food), Class I/II/III, against the official FDA
  openFDA enforcement reports — by product, firm or reason. Coverage US.
- input: query params `query` (required) — product/firm/reason, e.g. `metformin`; `category`
  (optional) — drug|device|food (default drug); `classification` (optional) — I|II|III;
  `limit` (optional, default 20, max 1000 — ALWAYS pass it, openFDA returns 1 by default).
- output: JSON. Example: {"total":94,"count":20,"recalls":[{"recall_number":"D-0435-2021",
  "classification":"Class II","status":"Terminated","reason_for_recall":"CGMP Deviations...",
  "recalling_firm":"Cardinal Health Inc.","recall_initiation_date":"20210315"}],"source":"openFDA..."}
- errors: 400 (query<2 / category|classification invalids), no-match → clean empty (count 0), 502/504.
- price: $0.01 (tier verif)
- source: openFDA — FDA enforcement reports (api.fda.gov) — public, no key
  (optional free openFDA key to raise the rate to 240 req/min)
- health: GET /recalls/health
- llm_usage_prompt: Product recall lookup against the official FDA openFDA enforcement reports (Recall Enterprise System). Search drugs, devices or food recalls by product, firm or reason, filter by Class I/II/III. Returns recall number, classification, status, reason, recalling firm, dates and distribution. Coverage US. For e-commerce, supply-chain, pharmacovigilance and compliance agents.

### GET /drug/label
- what: Drug interactions, warnings and contraindications lookup from the official FDA openFDA
  drug labels — by generic or brand name. Returns the key label sections. Coverage US.
- input: query params `drug` (required) — generic or brand name, e.g. `ibuprofen`; `limit`
  (optional, default 1, max 10).
- output: JSON. Example: {"drug":"ibuprofen","count":1,"labels":[{"generic_name":["IBUPROFEN"],
  "drug_interactions":"Ask a doctor before use if taking aspirin...","warnings":"Allergy alert...",
  "contraindications":null,"indications_and_usage":"temporarily relieves minor aches..."}],"source":"openFDA..."}
- errors: 400 (drug<2), 404 (no record), 502/504 (source).
- price: $0.05 (tier verdict)
- source: openFDA — FDA drug labels (api.fda.gov) — public, no key
- health: GET /drug/health
- llm_usage_prompt: Drug label lookup against the official FDA openFDA drug labels. By generic or brand name, returns key label sections: drug interactions, warnings, contraindications, indications and boxed warning. Coverage US. For healthcare, pharmacy and clinical-decision-support agents. Not medical advice.

### GET /osm/building-footprint
- what: Building footprint and geometry by coordinate from OpenStreetMap via the Overpass API.
  Returns each building as a GeoJSON polygon with server-computed footprint area (m²), useful
  tags (type, levels, height, name) and centroid. Worldwide coverage.
- input: query params (point or bbox) `lat` + `lon` — point, e.g. lat=48.8584 lon=2.2945;
  `radius` (optional) — metres [1-200], default 50; OR `bbox` — `south,west,north,east`
  (max ~2km per side). Bornes obligatoires (usage Overpass responsable).
- output: JSON. Example: {"query":{"lat":48.8584,"lon":2.2945,"radius_m":60},"count":1,
  "buildings":[{"osm_type":"way","osm_id":5013364,"tags":{"building":"tower","name":"Tour Eiffel",
  "height":"330"},"centroid":{"lat":48.8582,"lon":2.2944},"footprint_area_m2":9754.7,
  "geometry":{"type":"Polygon","coordinates":[[...]]}}],"source":"OpenStreetMap via Overpass",
  "attribution":"© OpenStreetMap contributors (ODbL)"}
- errors: 400 (neither coords nor bbox / radius out of bounds / bbox too large or malformed),
  clean empty (count 0) if no building, 503 (Overpass overloaded 429/504), 502/504 (source).
- price: $0.01 (tier verif)
- source: OpenStreetMap via Overpass API (overpass-api.de) — public, no key (bounded use)
- health: GET /osm/health
- llm_usage_prompt: Building footprint and geometry by coordinate via OpenStreetMap/Overpass. Returns the building polygon as GeoJSON, server-computed footprint area (m²), levels/floors, height and name for buildings around a point or within a bounding box. Worldwide. For real-estate, insurance, solar, geospatial and urban-planning agents. Input: lat/lon (+radius) or bbox.

### GET /climate-risk/score
- what: Derived climate risk score (0-100) by coordinate, computed server-side from Open-Meteo
  CMIP6 (HighResMIP downscaled) daily projections. Composite of heat, drought and extreme-rain
  sub-hazards via a documented deterministic formula, comparing a baseline window to a future
  horizon. Worldwide. DERIVED/HEURISTIC — not an official or regulatory risk rating.
- input: query params `lat` (required), `lon` (required); `scenario` (optional) — ssp245|ssp585
  (informational: the underlying CMIP6 HighResMIP dataset is a single high-emissions pathway and
  does not differentiate SSP); `horizon` (optional) — target year [2030-2050], default 2050.
- output: JSON. Example: {"composite_score":24.2,"rating":"Low","subscores":{"heat":16.5,
  "drought":43.2,"extreme_rain":15.4},"indicators":{...},"baseline":{"window":"2005-2014",
  "composite_score":18.0,...},"delta_vs_baseline":6.2,"scenario":"ssp245","horizon":2050,
  "methodology":{"summary":"0.45*heat+0.30*drought+0.25*extreme_rain ...","scenario_note":"..."},
  "source":"Open-Meteo Climate API — CMIP6 HighResMIP","attribution":"... (CC BY 4.0)"}
- errors: 400 (lat/lon missing ou out of bounds / scenario invalid / horizon out of bounds),
  404 (no CMIP6 data for these coordinates), 502/504 (source).
- price: $0.05 (verdict tier — value-added computed data)
- source: Open-Meteo Climate API (climate-api.open-meteo.com) — CMIP6 HighResMIP, CC BY 4.0, no key
- health: GET /climate-risk/health
- llm_usage_prompt: Derived climate risk score by coordinate from Open-Meteo CMIP6 (HighResMIP) projections. Returns a composite 0-100 heat/drought/extreme-rain risk score and rating, per-hazard subscores and indicators, baseline-vs-future comparison and an explicit methodology. Deterministic. Derived/heuristic, NOT an official or regulatory rating; the underlying dataset is a single high-emissions pathway and does not differentiate SSP scenarios. For property due-diligence, ESG, parametric-insurance and siting agents. Input: lat/lon (+scenario, horizon).

### GET /polymarket/resolution
- what: Polymarket market resolution oracle via the Gamma API. Returns whether a market is
  closed/resolved, the winning outcome (derived honestly from final outcome prices, only when
  unambiguous), resolution source, end date and condition id/slug — or lists recent resolved
  markets. DISTINCT from /polymarket/odds (live odds): this is the resolution state-of-truth.
- input: query params (optional) `slug` — market slug, e.g. `will-donald-trump-win-the-2024-us-
  presidential-election`; `condition_id` — on-chain condition id (0x…); `limit` — max markets
  when listing recent resolved markets [1-50], default 20. No slug/condition → resolved listing.
- output: JSON. Example: {"mode":"lookup","market":{"question":"...","closed":true,
  "resolution_status":"resolved","winning_outcome":"Yes","outcomes":[{"outcome":"Yes","price":1.0},
  {"outcome":"No","price":0.0}],"resolution_source":null,"end_date":"2024-11-05T12:00:00Z",
  "condition_id":"0x..."},"source":"Polymarket Gamma API"}
- errors: 400 (slug/condition_id malformed / limit out of bounds), 404 (market not found),
  503 (Gamma rate-limited 429), 502/504 (source).
- price: $0.01 (tier verif)
- source: Polymarket Gamma API (gamma-api.polymarket.com) — read-only, on-chain indexed, no auth
- health: GET /polymarket/resolution/health
- llm_usage_prompt: Polymarket market resolution oracle via the Gamma API. Returns whether a market is closed/resolved, the winning outcome (derived from final outcome prices, only when unambiguous — never invented), resolution source, end date and condition id/slug — or lists recent resolved markets. Distinct from live odds. For settlement, audit, research and event-forecasting agents. Input: market slug or condition id.

### GET /solar/building-insights
- what: Rooftop solar potential by coordinate from the official Google Maps Platform Solar API
  (buildingInsights:findClosest). Returns a clean synthesis: max panel count, usable array area
  (m²), annual sunshine hours, carbon-offset factor and whole-roof area for the closest building.
- input: query params `lat` (required), `lon` (required) — coordinate, e.g. lat=48.139 lon=11.566;
  `quality` (optional) — HIGH|MEDIUM|BASE (default HIGH; lower it to widen coverage).
- output: JSON. Example: {"center":{"latitude":48.1388,"longitude":11.5658},"imagery_quality":"HIGH",
  "postal_code":"80331","region_code":"DE","solar_potential":{"max_array_panels_count":48,
  "max_array_area_meters2":94.25,"max_sunshine_hours_per_year":1222.3,"carbon_offset_factor_kg_per_mwh":475.0,
  "whole_roof_area_meters2":194.63},"source":"Google Maps Platform Solar API","cached":false}
- errors: 400 (lat/lon missing or out of bounds / quality invalid), 404 (no building covered at
  this coordinate — retry quality=MEDIUM/BASE), 503 (quota), 502/504 (source).
- price: $0.05 (proprietary high-value data)
- source: Google Maps Platform Solar API (solar.googleapis.com) — API key required (in .env)
- health: GET /solar/health
- llm_usage_prompt: Rooftop solar potential by coordinate via the official Google Maps Platform Solar API. Returns max solar panel count, usable array area (m²), annual sunshine hours, carbon-offset factor (kg/MWh) and whole-roof area for the closest building. Worldwide partial coverage (EU well covered). For agents sizing solar installs, real-estate energy scoring, or carbon/ROI estimates. Input: lat/lon (+imagery quality).

### GET /uk-companies/search
- what: Search the official UK Companies House register by name or number, or fetch a detailed
  company profile by number (registered address, SIC codes, officers). KYB lookup, UK coverage.
- input: query params `q` (required unless company_number) — name or number, e.g. `Tesco`;
  `company_number` (optional) — if set, returns the detailed profile via /company/{number};
  `limit` (optional, default 20, max 50); `start_index` (optional, pagination).
- output: JSON. Search example: {"mode":"search","total_results":356,"count":1,"companies":[{"title":
  "TESCO PLC","company_number":"00445790","company_status":"active","company_type":"plc",
  "date_of_creation":"1947-11-27","address_snippet":"Tesco House, ... AL7 1GA"}]}. Profile mode adds
  registered_office_address, sic_codes, has_insolvency_history and officers (with
  identity_verification_status when present, post-ECCTA).
- errors: 400 (q < 2 chars / company_number invalid / limit out of bounds), 404 (unknown number),
  503 (rate-limit), 502/504 (source).
- price: $0.01 (lookup registre officiel)
- source: UK Companies House Public Data API (api.company-information.service.gov.uk) — Basic auth, key in .env
- health: GET /uk-companies/health
- llm_usage_prompt: UK company verification via the official Companies House register. Search by name or number, or fetch a detailed profile by company number: status, type, incorporation date, registered office address, SIC codes, insolvency history and officers (with post-ECCTA identity verification status when present). For KYB, AML and supplier due-diligence agents. Input: company name or number.

### GET /patents/search
- what: European patent search and bibliographic lookup against the official EPO Open Patent
  Services (OPS v3.2). Lookup by publication number, or CQL search by title, applicant, inventor
  and publication-date range. Returns parsed bibliographic data per patent.
- input: query params (at least one of) `publication_number` — e.g. `EP1000000` (→ biblio);
  `title` (CQL ti=); `applicant` (CQL pa=); `inventor` (CQL in=); `date_from`/`date_to`
  (ISO YYYY-MM-DD); `limit` (optional, default 20, max 50).
- output: JSON. Example: {"mode":"publication","count":1,"patents":[{"publication_number":"EP1000000A1",
  "country":"EP","kind_code":"A1","publication_date":"20000517","family_id":"19768124","titles":
  {"en":"Apparatus for manufacturing green bricks...","de":"..."},"applicants":["BEHEERMIJ DE BOER..."],
  "inventors":["KOSMAN WILHELMUS..."],"ipc_classifications":["B28B1/29"],"cpc_classifications":["B28B1/29"]}]}
- errors: 400 (no criteria / CQL invalid / non-ISO date), 404 (unknown number),
  503 (EPO fair-use quota exceeded), 502/504 (source). OAuth2 token transparently renewed.
- price: $0.05 (structured institutional data, value-added parsing)
- source: EPO Open Patent Services v3.2 (ops.epo.org) — OAuth2 client_credentials, keys in .env
- health: GET /patents/health (verifies the token, does not consume the data quota)
- llm_usage_prompt: European patent search and bibliographic lookup via the official EPO Open Patent Services (OPS). By publication number, or CQL search by title, applicant, inventor and publication-date range. Returns publication number, country, kind code, dates, multilingual titles, applicants, inventors, IPC/CPC classifications and INPADOC family id. Coverage EU and worldwide (100+ patent offices). For IP due-diligence, freedom-to-operate, prior-art and R&D agents. Input: patent number or search criteria.

### GET /fr-sirene/lookup
- what: French company/establishment lookup against the official INSEE Sirene register — by SIRET
  (establishment), SIREN (legal unit) or free-text/multi-criteria search.
- input: query params (at least one of) `siret` — 14 digits, e.g. `44306184100047`; `siren` —
  9 digits, e.g. `443061841`; `q` — free-text/multi-criteria, e.g. `denominationUniteLegale:"GOOGLE FRANCE"`;
  `limit` (optional, default 20, max 100).
- output: JSON. Example: {"mode":"siret","etablissement":{"siren":"443061841","siret":"44306184100047",
  "denomination":"GOOGLE FRANCE","etat_administratif":"A","date_creation":"2011-05-13",
  "activite_principale_naf":"62.02A","tranche_effectifs":"42","statut_diffusion":"O",
  "adresse":"8 RUE DE LONDRES 75009 PARIS"},"source":"INSEE Sirene V3.11 (api.insee.fr)"}
- errors: 400 (no param / siret≠14 / siren≠9 / limit), 404 (SIRET/SIREN unknown),
  502 (invalid key → 401 source), 503 (rate-limit 30/min), 502/504 (source).
- price: $0.01 (lookup registre officiel)
- source: INSEE Sirene V3.11 (api.insee.fr) — API key in header X-INSEE-Api-Key-Integration (in .env)
- health: GET /fr-sirene/health
- llm_usage_prompt: French company/establishment verification via the official INSEE Sirene register. Lookup by SIRET, SIREN or name: legal name, administrative status (active/closed), legal form, NAF/APE activity code, creation date, headcount band and registered address. For KYB, AML and supplier due-diligence agents (France). Input: SIRET, SIREN or company name.

### GET /iban/validate
- what: IBAN format validation via the official ISO 13616 mod-97 algorithm — 100% local and
  deterministic, no external source, no key. Structural validation + country/bank-code extraction.
- input: query param `iban` (required), e.g. `FR1420041010050500013M02606`.
- output: JSON. Example: {"iban":"FR14...","country_code":"FR","check_digits":"14","formatted":
  "FR14 2004 1010 0505 0001 3M02 606","valid":true,"reason":null,"bank_code":"20041",
  "source":"ISO 13616 mod-97 local validation (no external source)"}. Invalid → valid:false + reason
  (length_mismatch / unknown_country / checksum_failed / country_code_invalid).
- errors: 400 (iban missing / aberrant length / non-alphanumeric characters).
- price: $0.01 (instant local validation)
- source: ISO 13616 mod-97, local (no external source, no key) — FORMAT check, NOT account existence
- health: GET /iban/health
- llm_usage_prompt: IBAN format validation via the ISO 13616 mod-97 algorithm (local, deterministic). SEPA/EU structural validation: returns valid/invalid with a clear reason, country code, check digits, bank code and grouped formatting. Format check only — does NOT verify the account exists. For payment, onboarding and treasury agents. Input: IBAN.

### GET /fr-legifrance/search
- what: French legal text search and lookup against the official Légifrance (DILA, via the PISTE
  portal) — keyword search across codes, laws/decrees (LODA) and the Journal Officiel (JORF), or
  consult a text by id. Complements EUR-Lex (EU law) on the national French side.
- input: query params `q` — keyword, e.g. `données personnelles`; `fond` (optional) —
  LODA_DATE | CODE_DATE | JORF (default LODA_DATE); `text_id` (optional) — consult a text by id/cid;
  `limit` (optional, default 10, max 50).
- output: JSON. Example: {"mode":"search","query":"données personnelles","fond":"LODA_DATE","total":107757,
  "count":1,"results":[{"id":"LEGITEXT000048853940_01-01-2024","cid":"JORFTEXT000048736062","title":"Décret n° 2023-1397
  du 29 décembre 2023 relatif au traitement des données personnelles...","nature":"DECRET","etat":"VIGUEUR"}],
  "source":"Légifrance via DILA/PISTE"}
- errors: 400 (no criteria / fond invalid), 404 (text unknown), 502 (PISTE auth rejected /
  source), 503 (quota PISTE). OAuth2 token transparently renewed.
- price: $0.05 (structured legal data)
- source: Légifrance via DILA/PISTE (api.piste.gouv.fr) — OAuth2 client_credentials, keys in .env
- health: GET /fr-legifrance/health (verifies the token, does not run a search)
- llm_usage_prompt: French law search and lookup via the official Légifrance (DILA/PISTE). Keyword search across codes, laws/decrees (LODA) and the Journal Officiel (JORF), or consult a text by id. Returns title, identifier (id/cid), nature, date and legal status (in force/repealed). Complements EUR-Lex on the national French side. For legal-research, regulatory and compliance agents. Input: keyword (+corpus) or text id.

### GET /ecb/exchange-rate
- what: Euro foreign-exchange reference rate lookup against the official European Central Bank (ECB)
  SDMX data portal — latest or by date, for any of the ~29 published currencies.
- input: query params `currency` (required) — 3-letter ISO, e.g. `USD`, `GBP`, `CHF`, `JPY`;
  `frequency` (optional) — D (daily, default) or M (monthly); `date` (optional) — ISO YYYY-MM-DD,
  default = latest available.
- output: JSON. Example: {"currency":"USD","base":"EUR","rate":1.134,"observation_date":"2026-06-24",
  "frequency":"D","series_key":"D.USD.EUR.SP00.A","source":"European Central Bank — euro reference rates"}
- errors: 400 (currency≠3 lettres / EUR / frequency / date non-ISO), 404 (devise inconnue ou pas de
  rate on this date — non-business day), 502/504 (source).
- price: $0.01 (official public data lookup)
- source: European Central Bank SDMX 2.1 (data-api.ecb.europa.eu) — public, no key. Reference rates, not live market.
- health: GET /ecb/health
- llm_usage_prompt: Euro foreign-exchange reference rate lookup via the official European Central Bank (ECB) SDMX data portal. Returns the EUR reference rate for a currency (29 currencies), latest or by date, with the real observation date. Reference rates (published ~16:00 CET on working days), not live market rates. For pricing, accounting, FX and treasury agents. Input: currency (+frequency, date).

### GET /crypto/token-safety
- what: Token safety check before buying — is this token a honeypot or a rug pull? Bundles honeypot detection, buy/sell tax, holder concentration, LP lock and liquidity (GoPlus, Honeypot.is, DexScreener) into a single 0-100 token security risk score with a clear buy/avoid verdict. One call replaces three lookups. EVM chains.
- price: $0.05
- input:
  - token: Token contract address (0x + 40 hex) (required)
  - chain: base | ethereum | bsc | polygon | arbitrum | optimism | avalanche (default base)
- source: GoPlus Security + Honeypot.is + DexScreener
- health: GET /crypto/token-safety/health
- llm_usage_prompt: Token safety check before buying: is this token a honeypot or rug pull? Honeypot, buy/sell tax, holder concentration, LP lock and liquidity bundled into a 0-100 token security risk score with a buy/avoid verdict. For trading agents screening a token. Input: token contract address + chain.

### GET /crypto/derivatives-radar
- what: Perpetual funding rates, open interest and long/short ratios across Binance, Bybit, OKX and Hyperliquid in one call, plus the computed funding-rate arbitrage spread between venues and a crowding signal. Multi-venue perp derivatives data, resilient: returns whatever venues respond. For trading and funding-arbitrage agents.
- price: $0.05
- input:
  - symbol: Coin ticker, e.g. 'BTC', 'ETH', 'SOL'
- source: Binance + Bybit + OKX + Hyperliquid (perp public APIs)
- health: GET /crypto/derivatives-radar/health
- llm_usage_prompt: Perpetual funding rates, open interest and long/short ratios across Binance/Bybit/OKX/Hyperliquid in one call, plus the computed funding-rate arbitrage spread and crowding signal. For trading and funding-arb agents. Input: coin ticker.

### GET /crypto/wallet-xray
- what: Wallet risk analysis in one call — native + ERC-20 balances, USD portfolio valuation, token count and an OFAC sanctions flag (Blockscout, DexScreener, OFAC list). A full on-chain wallet x-ray with risk flags, where primitives serve one field per call this returns the whole valued portfolio plus risk. EVM chains.
- price: $0.05
- input:
  - address: Wallet address (0x + 40 hex) (required)
  - chain: base | ethereum | optimism | arbitrum | polygon | gnosis (default base)
- source: Blockscout + DexScreener + OFAC sanctioned-address list
- health: GET /crypto/wallet-xray/health
- llm_usage_prompt: Wallet risk analysis: native + ERC-20 balances, USD portfolio valuation, token count and an OFAC sanctions flag, bundled in one call. For on-chain analysis, compliance and trading agents. Input: wallet address + chain.

### GET /crypto/dex-cex-spread
- what: DEX vs CEX arbitrage spread — the price difference between DEX and CEX with the fee-adjusted NET spread and execution direction, not two raw prices. Compares the most liquid DEX pair (DexScreener) to CEX spot (Binance/Bybit), subtracts taker+swap fees, and flags any profitable arbitrage opportunity. For arbitrage agents.
- price: $0.05
- input:
  - symbol: Coin ticker, e.g. 'ETH', 'WBTC', 'ARB'
- source: DexScreener (DEX) + Binance/Bybit spot (CEX)
- health: GET /crypto/dex-cex-spread/health
- llm_usage_prompt: DEX vs CEX arbitrage spread: the price difference between DEX and CEX with the fee-adjusted NET spread (after taker+swap fees) and execution direction, plus whether the arbitrage is profitable. For arbitrage agents. Input: coin ticker.

### GET /web/extract
- what: Web search and webpage extraction for AI agents — returns clean, LLM-ready markdown (deduplicated, with source URLs), not raw HTML. Give 'query' to search the web or 'url' to scrape and extract a page into markdown ready to drop into a prompt. Powered by Jina Reader.
- price: $0.05
- input:
  - query: Web search query
  - url: If set, scrape and extract this page as clean markdown instead of searching
  - limit: Max search results [1-10]
- source: Jina AI Reader (r.jina.ai) + Search (s.jina.ai)
- health: GET /web/extract/health
- llm_usage_prompt: Web search and webpage extraction for AI agents, returned as clean LLM-ready markdown with source URLs. Give 'query' to search the web or 'url' to scrape and extract a page. For research and RAG agents.

### GET /crypto/new-pairs
- what: Newly launched DEX pairs and pools per chain (GeckoTerminal) with an instant per-pair safety check bundled in (GoPlus honeypot/tax/verified + mini score) — a sniping agent filters rugs without a second call. Returns pool, base token, price, liquidity and age. EVM safety; Solana lists pairs only.
- price: $0.05
- input:
  - chain: base | ethereum | bsc | polygon | arbitrum | optimism | avalanche | solana
  - limit: Max new pairs [1-15]
  - safety: Bundle a GoPlus safety check per pair (EVM only), default true
- source: GeckoTerminal (new pools) + GoPlus Security (token safety)
- health: GET /crypto/new-pairs/health
- llm_usage_prompt: Newly launched DEX pairs and pools per chain with an instant per-pair safety check bundled in (honeypot/tax/verified + mini score), so a sniping agent filters rugs without a second call. Input: chain + limit.

### GET /x402/seller-trust
- what: Score an x402 seller wallet BEFORE paying it: verdict TRUSTED/CAUTION/AVOID + confidence from its on-chain settlement graph on Base — settlement count, unique paying counterparties, wallet age, wash-trade and sybil signals — with a signed, offline-verifiable receipt. Vet / due-diligence an unknown x402 endpoint or merchant wallet before you pay it: is this seller a scam, a fake-volume wash-trader, or a sybil cluster? Agent counterparty risk, seller reputation and trust score for x402 / agentic payments.
- price: $0.01
- input:
  - seller_wallet: x402 seller payTo wallet (0x + 40 hex)
  - depth: 'shallow' (~200 settlements) or 'deep' (~500)
- source: Base on-chain settlement graph (Blockscout) + OFAC SDN crypto list
- health: GET /x402/seller-trust/health
- llm_usage_prompt: Score an x402 seller wallet TRUSTED/CAUTION/AVOID before paying it, from its on-chain settlement graph (settlements, counterparties, age, wash-trade/sybil) with a signed offline-verifiable receipt. Input: seller_wallet.

### GET /x402/payment-firewall
- what: Pre-payment risk firewall for AI agents: validate {recipient, amount, chain} BEFORE sending an x402 payment and get ALLOW/REVIEW/BLOCK in <200ms — OFAC blocklist, sanctioned-mixer check, runaway-overspend vs expected price and anomalous-amount guard. Stop an agent from overpaying or paying a bad address; transaction risk screening and payment guardrail before send. The hosted spending guard / agent budget protection an offline SDK cannot give.
- price: $0.005
- input:
  - to_address: Recipient/payTo address (EVM 0x.., Solana, BTC, TRON, XRP)
  - amount: Payment amount in USD, e.g. 0.01
  - chain: base | ethereum | solana | bitcoin | tron (optional; inferred from address)
  - expected_price: Price the resource advertised — flags runaway overspend
- source: OFAC SDN crypto blocklist (0xB10C mirror) + chain validation + spend baseline
- health: GET /x402/payment-firewall/health
- llm_usage_prompt: Pre-payment firewall: validate {recipient, amount, chain} before an x402 payment, get ALLOW/REVIEW/BLOCK in <200ms (OFAC blocklist, mixer, runaway-overspend). Input: to_address + amount.

### GET /compliance/wallet-screen
- what: Crypto wallet sanctions & compliance screening (AML/KYT): screen any EVM/Solana/Bitcoin/Tron/XRP wallet against the official OFAC SDN crypto list (UN/UK declared) with mixer (Tornado Cash) exposure and wallet age, returning PASS/WARN/BLOCK and a cryptographically SIGNED compliance receipt (list-version-pinned, verifiable offline) for audit. Is this address sanctioned, blacklisted or mixer-linked? OFAC wallet check, travel-rule / KYT counterparty screening with an auditable attestation an agent can keep.
- price: $0.005
- input:
  - wallet: Wallet to screen — EVM '0x..', Solana, BTC, TRON 'T..', XRP 'r..'
  - chains: Optional comma-separated EVM chains for exposure/age, e.g. 'ethereum,base'
- source: Official OFAC SDN crypto list + Tornado Cash mixer labels + Blockscout
- health: GET /compliance/wallet-screen/health
- llm_usage_prompt: Crypto wallet sanctions/AML screening (OFAC SDN multi-chain) with mixer exposure, wallet age and a signed compliance receipt for audit. Verdict PASS/WARN/BLOCK. Input: wallet.

### GET /macro/snapshot
- what: Macro economic snapshot API — dated Eurozone economic indicators for trading/research agents: euro-area / Eurozone inflation (HICP / CPI), ECB policy interest rates (deposit, refi), unemployment and EUR FX from the official European Central Bank data portal — each indicator with value, as_of date, next_release cadence and a signed dated snapshot receipt for reproducible backtests. The interest-rate, inflation and currency context an agent needs before trading. Central bank / economic indicators data API.
- price: $0.005
- input:
  - indicators: Comma-separated: inflation_hicp, core_inflation, unemployment, deposit_facility_rate, main_refi_rate, marginal_lending_rate, fx_usd, fx_gbp, fx_jpy, fx_chf
  - area: HICP reference area: 'U2' (euro area, default) or a country code e.g. 'FR','DE'
- source: European Central Bank — Data Portal SDMX 2.1 (data-api.ecb.europa.eu)
- health: GET /macro/snapshot/health
- llm_usage_prompt: Dated euro-area macro snapshot (HICP inflation, ECB policy rates, unemployment, EUR FX) from the official ECB data portal, with next_release and a signed dated snapshot receipt for backtests. Input: indicators.

### GET /solana/token-safety
- what: Solana SPL token safety, rug check and honeypot / scam detection before trading: is this Solana token a rug pull, honeypot or scam? Verdict SAFE/RISKY/CRITICAL + 0-100 score combining STATIC checks (mint/freeze authority, holder concentration) AND BEHAVIORAL analysis (liquidity adequacy, churn, recent dump, tx velocity) plus a blue-chip false-positive guard so USDC/USDT/SOL are never flagged as rugs. Pre-trade SPL token security / scam-token detection that catches behavioral rugs static checkers miss.
- price: $0.01
- input:
  - mint: SPL token mint address (base58), e.g. 'EPjFW...Dt1v' (USDC)
  - deep: Deeper behavioral + holder analysis (more RPC calls)
- source: Solana RPC (public) + DexScreener
- health: GET /solana/token-safety/health
- llm_usage_prompt: Solana SPL token rug check SAFE/RISKY/CRITICAL: static (mint/freeze authority, holders) + behavioral (liquidity, churn, dump, velocity) + blue-chip false-positive guard. Input: mint.

### GET /solana/pre-trade
- what: All-in-one Solana pre-trade decision in ONE call: BUY-SAFE/CAUTION/AVOID fusing four scored modules — token security, EXECUTABLE liquidity depth (estimated slippage at $100/$1k/$10k), deployer history/control and holder concentration. Should I buy or avoid this Solana token? Full token due-diligence in one call — replaces 3-4 lookups; built for a trading / sniping agent's risk-review pipeline. Solana trading safety decision and buy/avoid verdict.
- price: $0.05
- input:
  - mint: SPL token mint address (base58) to evaluate before buying
- source: Solana RPC (public) + DexScreener
- health: GET /solana/pre-trade/health
- llm_usage_prompt: One-call Solana pre-trade decision BUY-SAFE/CAUTION/AVOID fusing token security + executable liquidity depth (slippage at size) + deployer history + holder concentration. Input: mint.

### GET /crypto/signal-fusion
- what: Crypto trading signal (long/short) and market-regime signal in ONE call: LONG/SHORT/NEUTRAL with confidence, fusing the cross-exchange funding rate, crowding (long/short) bias, trend/chop market regime and BTC→altcoin lead-lag from Binance/Bybit/OKX/Hyperliquid — with declared data freshness. Should I go long or short on this coin? A fused directional trading signal where competitors sell the funding rate, regime and the four numbers separately. Derivatives trading signal for any altcoin.
- price: $0.02
- input:
  - symbol: Coin ticker, e.g. 'BTC','ETH','SOL'
  - timeframe: Regime/lead-lag timeframe: 5m | 15m | 1h | 4h (default 1h)
- source: Binance + Bybit + OKX + Hyperliquid (perp + klines)
- health: GET /crypto/signal-fusion/health
- llm_usage_prompt: One-call fused trading signal LONG/SHORT/NEUTRAL combining cross-exchange funding, crowding, trend/chop regime and BTC→alt lead-lag. Input: symbol + timeframe.

### GET /agent/due-diligence
- what: Vet any company, crypto project, counterparty or domain before engaging: one call composes a risk DOSSIER — synthesis + detected signals + sources + 0-100 risk score + GO/CAUTION/STOP verdict + recommendation + a signed, offline-verifiable receipt. Is this entity a scam, fraud, sanctioned, or safe to deal with? Replaces 3-4 manual web searches plus the synthesis an agent would have to do itself, composed from live public web signals. Counterparty / vendor / project due-diligence for agents.
- price: $0.50
- input:
  - target (required): Entity / project / counterparty / domain to vet, e.g. 'Acme DeFi Labs'
  - target_type: Optional: entity | project | counterparty | domain | wallet
  - context: Optional: why you're engaging / what you're about to do
- source: Public web signals (Jina Reader, keyless) + Claude Haiku synthesis
- health: GET /agent/due-diligence/health
- llm_usage_prompt: Composed risk dossier on any entity/project/counterparty/domain: summary + signals + sources + 0-100 score + GO/CAUTION/STOP + signed receipt. One call replaces 3-4 lookups. Input: target.

### GET /agent/analysis-report
- what: Structured analysis report on any URL, company, product or subject in one call: positioning + strengths + risks + opportunities + 0-100 actionable score + STRONG/MODERATE/WEAK rating + recommendation + provenance + a signed receipt. A fixed-schema research report an agent can consume without re-parsing, composed from live public web signals. Competitive / investment / partnership analysis for agents.
- price: $0.25
- input:
  - input (required): URL, company, product or subject, e.g. 'stripe.com'
  - focus: Optional: investment | competitive | partnership | security | general
- source: Public web signals (Jina Reader, keyless) + Claude Haiku synthesis
- health: GET /agent/analysis-report/health
- llm_usage_prompt: Fixed-schema analysis report on a URL/company/subject: positioning + strengths + risks + opportunities + 0-100 score + STRONG/MODERATE/WEAK + signed receipt. Input: input (+ optional focus).

### GET /agent/clearance
- what: Safety preflight before an agent runs a DESTRUCTIVE, costly or IRREVERSIBLE action (delete, deploy, transfer funds, publish, overwrite): risk assessment + reversibility check + timing + CLEAR/LIMIT/REVIEW/BLOCK verdict + a cryptographically SIGNED ClearancePacket (decision, approved amount, policy version, evidence hash, expiry, rationale) the agent keeps as proof of authorization. Should I run this action, or stop and ask a human? Stops a mistake far costlier than the call price. Agent action guardrail / clearance / human-in-the-loop gate.
- price: $0.25
- input:
  - action (required): Action about to run, e.g. 'delete the production users table'
  - action_type: Optional: delete|deploy|send|spend|transfer|publish|email|overwrite|approve|rotate|shutdown|other
  - amount_usd: Monetary stake in USD if applicable, e.g. 250
  - reversible: Optional caller hint: is the action reversible?
  - target: Optional target (table, address, repo, recipient)
  - context: Optional extra context (environment, prior approval)
- source: Local deterministic clearance rules engine + Ed25519 signed ClearancePacket
- health: GET /agent/clearance/health
- llm_usage_prompt: Preflight for a destructive/costly/irreversible agent action: risk + reversibility + timing + CLEAR/LIMIT/REVIEW/BLOCK + signed ClearancePacket. Input: action (+ action_type, amount_usd).

### GET /agent/output-qa
- what: Review and improve an agent's outbound text (email, social post, customer reply) before sending: a multi-criteria scorecard (clarity, spam-safety, tone, length, personalization, CTA, compliance — each 0-100) + poor/fair/good/excellent rating + top suggestions + a ready-to-send IMPROVED REWRITE. Will this message land or get flagged as spam? One call scores and rewrites. Output review / spam check / copy QA for agents.
- price: $0.10
- input:
  - output (required): The agent's outbound text to review (email, post, reply)
  - format: Optional: email | social_post | customer_reply | other
  - goal: Optional intended goal, e.g. 'book a demo'
- source: Claude Haiku rubric scoring + rewrite
- health: GET /agent/output-qa/health
- llm_usage_prompt: Score an agent's outbound text on 7 criteria (clarity, spam-safety, tone, length, personalization, CTA, compliance) + rating + suggestions + improved rewrite. Input: output (+ format, goal).

### GET /agent/content-scan
- what: Scan any external content before an agent INGESTS it (SKILL.md, web page, user input, tool output) for prompt-injection, data exfiltration, dangerous code execution, hidden/invisible unicode and instruction overrides: 0-100 risk score + SAFE/WARN/BLOCK verdict + structured findings with matched evidence. Is it safe to feed this content to my agent? A fast, cheap, deterministic, high-volume pre-ingest firewall. Prompt-injection / skill-audit / untrusted-content security check for agents.
- price: $0.10
- input:
  - content (required): Raw content to scan before ingestion (SKILL.md, web page, user input, tool output)
  - source_type: Optional: skill | webpage | user_input | tool_output | document
- source: Local deterministic pattern engine (regex + unicode heuristics)
- health: GET /agent/content-scan/health
- llm_usage_prompt: Deterministic pre-ingest scan for prompt-injection, exfiltration, dangerous exec, hidden unicode and instruction overrides: 0-100 risk score + SAFE/WARN/BLOCK + structured findings. Input: content.

### GET /crypto/pre-trade-verdict
- what: One-call GO/CAUTION/NO-GO pre-trade verdict for AI trading agents: fuses token safety (honeypot, rug, tax, holders), counterparty wallet sanctions screening (OFAC/mixer) and cross-exchange market signal into a single decision with a signed, offline-verifiable receipt. Should I buy this token now? Replaces three separate calls (token-safety + wallet-screen + signal) with one fused GO/NO-GO verdict. EVM chains and Solana.
- price: $0.05
- input:
  - token: Token contract (EVM 0x+40hex) or SPL mint (base58) to evaluate
  - chain: base | ethereum | bsc | polygon | arbitrum | optimism | avalanche | solana (default base)
  - wallet: Optional counterparty wallet to screen (OFAC/mixer)
- source: Internal token-safety + wallet-screen + signal-fusion (composed) + Ed25519 receipt
- health: GET /crypto/pre-trade-verdict/health
- llm_usage_prompt: One-call GO/CAUTION/NO-GO pre-trade verdict fusing token safety (honeypot/rug/tax) + counterparty sanctions screening (OFAC/mixer) + cross-exchange market signal, with a signed receipt. Replaces three calls. EVM + Solana. Input: token (+chain, optional wallet).

### GET /agent/ledger
- what: Your AI agent's books in one call: structured USDC accounting for any wallet on Base (the x402 settlement asset) — revenue (incoming) vs expenses (outgoing), net, transaction count, top paying counterparties and payees, and a weekly breakdown over a 7d/30d/90d window. Turnkey bookkeeping / expense and revenue report an agent can drop straight into its accounting. The value is the structuring, not raw data access.
- price: $0.04
- input:
  - wallet: EVM wallet address (0x + 40 hex) to report on
  - period: Reporting window: 7d | 30d | 90d | all (default 30d)
- source: Base Blockscout USDC transfers (base.blockscout.com), settlement asset for x402
- health: GET /agent/ledger/health
- llm_usage_prompt: Structured USDC accounting for a wallet on Base: revenue vs expenses, net, tx count, top payers/payees, weekly breakdown. Your agent's books in one call. Input: wallet (+period 7d/30d/90d).

### GET /proof/notarize
- what: Notarize a document or data hash for AI agents: submit a SHA-256 hash (or short content hashed server-side) and get a timestamped, Ed25519-signed proof-of-existence receipt, verifiable offline with no callback. Prove an artifact, agent output or decision existed at a point in time. Honest scope: this is a SIGNED proof-of-existence receipt, NOT a blockchain-anchored or eIDAS/RFC-3161 legal timestamp. The issuer never sees your content when you send only a hash.
- price: $0.01
- input:
  - hash: SHA-256 hex digest (64 hex chars) to notarize
  - content: Alternative to hash: short content hashed server-side (SHA-256), max 100000 chars
  - memo: Optional label bound into the signed proof, e.g. 'invoice #42'
- source: Local Ed25519 proof-of-existence (no external source)
- health: GET /proof/notarize/health
- llm_usage_prompt: Notarize a SHA-256 hash (or short content hashed server-side) into a timestamped, Ed25519-signed proof-of-existence receipt, verifiable offline. Signed proof, NOT a blockchain-anchored or legal (eIDAS/RFC 3161) timestamp. Input: hash or content (+memo).

### GET /crypto/token-dossier
- what: Full degen token dossier in one call: safety score + honeypot/tax + detailed TOP HOLDERS and concentration + liquidity/FDV/volume/pool-age + contract control (owner, creator, mintable, open-source) + an AI red-flag narrative. Deep due-diligence report on a token before aping — the premium tier above a plain safety check. EVM chains and Solana.
- price: $0.10
- input:
  - token: Token contract (EVM 0x+40hex) or SPL mint (base58)
  - chain: base | ethereum | bsc | polygon | arbitrum | optimism | avalanche | solana (default base)
- source: GoPlus Security + Honeypot.is + DexScreener (+ Solana RPC) + Claude Haiku (narrative)
- health: GET /crypto/token-dossier/health
- llm_usage_prompt: Full token dossier: safety score + top holders + concentration + liquidity/FDV/pool-age + contract control (owner/creator/mintable) + AI red-flag narrative. Premium tier above a plain safety check. EVM + Solana. Input: token (+chain).

### GET /market/intel
- what: AI market intelligence report in one call for research and trading agents: cross-exchange perp derivatives (funding, open interest, long/short), fused directional signals and the euro-area macro snapshot (ECB inflation, policy rates, unemployment, EUR FX) synthesized into a headline, crypto and macro outlook, key signals, risks and opportunities. Multi-source market research report with a crypto + macro/EU lens.
- price: $0.10
- input:
  - focus: crypto | macro | all (default all)
  - depth: quick | deep — deep adds SOL and more web signals (default quick)
- source: Binance/Bybit/OKX/Hyperliquid perp + ECB SDMX macro + public web (Jina) + Claude Haiku
- health: GET /market/intel/health
- llm_usage_prompt: AI market intelligence report: cross-exchange perp derivatives (funding/OI/long-short) + fused directional signals + euro-area ECB macro (inflation, rates, unemployment, FX) synthesized into headline, crypto and macro outlook, key signals, risks, opportunities. Input: focus (crypto/macro/all), depth.

### GET /crypto/wallet-forensics
- what: Multi-hop wallet forensics for AI agents: traverse an address's ERC-20 counterparty graph across 1-3 hops and get the flow graph (nodes + edges), ranked counterparties (direction, tokens, activity) and detected patterns — flow concentration, circular flows, linked/clustered wallets and high-degree hubs (exchange/mixer/deployer) — plus a signed receipt. Trace where funds move and who is connected. EVM chains.
- price: $0.15
- input:
  - wallet: EVM wallet address (0x + 40 hex) to investigate
  - depth: Graph hops to traverse: 1, 2 or 3 (default 2)
  - chain: base | ethereum | optimism | polygon | arbitrum | gnosis (default base)
- source: EVM Blockscout ERC-20 transfer graph (multi-hop) + OFAC SDN list + Ed25519 receipt
- health: GET /crypto/wallet-forensics/health
- llm_usage_prompt: Multi-hop wallet forensics: ERC-20 counterparty flow graph (1-3 hops), ranked counterparties, pattern detection (flow concentration, circular flows, linked wallets, high-degree hubs) + signed receipt. Input: wallet (+depth 1-3, chain).

### GET /chain/events
- what: On-demand, decoded on-chain events for AI agents: query a contract's event logs on Base (eth_getLogs) and get normalized, human-readable events — Transfer, Approval, ERC-1155 TransferSingle/Batch, Uniswap V2/V3 Swap, Sync, Mint, Burn — with raw topics for unknown signatures. Filter by event name, full signature or topic0, and by block range or lookback. Decoded event feed with no RPC setup, block span capped.
- price: $0.05
- input:
  - contract: Contract address to read logs from (0x + 40 hex)
  - event: Event name (Transfer/Approval/Swap/Mint/Burn/TransferSingle/Sync), full signature 'Transfer(address,address,uint256)', or 0x topic0. Omit for all.
  - from_block: Start block (inclusive). Omit to use lookback.
  - to_block: End block (inclusive). Omit for latest.
  - lookback: If from/to omitted, scan the last N blocks (1-5000, default 2000)
  - limit: Max events (1-100, most recent in range)
- source: Base JSON-RPC eth_getLogs (Alchemy, fallback public Base RPC)
- health: GET /chain/events/health
- llm_usage_prompt: Decoded, normalized on-chain events via eth_getLogs on Base: Transfer, Approval, ERC-1155 TransferSingle/Batch, Uniswap V2/V3 Swap, Sync, Mint, Burn — raw topics for unknown signatures. Filter by event name/signature/topic0 and block range. Input: contract (+event, block range/lookback).

### GET /x402/seller-audit
- what: Deep audit of an x402/ACP seller before trusting it: give a seller wallet or domain and get a full report — on-chain settlement reputation (settlements, unique counterparties, wash-trade/sybil, OFAC), the seller's advertised endpoints (from its /.well-known/x402.json), on-chain liveness, web reputation and red flags — fused into a TRUSTED/CAUTION/AVOID verdict with a signed, offline-verifiable receipt. Deeper than a pass/fail badge: the full seller due-diligence dossier for agentic payments.
- price: $0.50
- input:
  - seller: Seller wallet (0x + 40 hex) or domain, e.g. 'api.example.com'
  - depth: On-chain depth: 'shallow' (~200 settlements) or 'deep' (~500)
- source: Seller /.well-known/x402.json discovery + internal seller-trust + public web signals + Claude Haiku + Ed25519 receipt
- health: GET /x402/seller-audit/health
- llm_usage_prompt: Deep x402 seller audit: on-chain settlement reputation (wash-trade/sybil/OFAC) + advertised endpoints from its /.well-known/x402.json + on-chain liveness + web reputation fused into TRUSTED/CAUTION/AVOID with a signed receipt. Deeper than a pass/fail badge. Input: seller wallet or domain.

### GET /agent/clearance-packet
- what: Signed clearance packet an AI agent buys BEFORE a large capital move: one call fuses OFAC/sanctions + mixer screening, multi-hop fund-flow forensics (layering, concentration, hubs) and counterparty settlement trust into a CLEAR/REVIEW/BLOCK verdict with the evaluated amount, a policy-bounded approved amount, a short expiry and an Ed25519 signature verifiable offline. Can I send this capital to THIS wallet? Pre-payment AML/KYT clearance and counterparty due-diligence for agents.
- price: $1.50
- input:
  - to: Counterparty EVM wallet the capital will move to (0x + 40 hex)
  - amount_usd: Capital about to move, in USD, e.g. 800 — scopes the verdict and approved amount
  - depth: Forensics graph hops: 1, 2 or 3 (default 2)
- source: Internal wallet-screen (OFAC/mixer) + wallet-forensics (multi-hop) + seller-trust (settlement) composed + Ed25519 packet
- health: GET /agent/clearance-packet/health
- llm_usage_prompt: Signed clearance packet before an agent moves significant capital to a wallet: fuses OFAC/sanctions + mixer screening, multi-hop fund-flow forensics (circular/layering, concentration, hubs) and on-chain counterparty settlement trust into one CLEAR/REVIEW/BLOCK verdict with the evaluated amount, a policy-bounded approved amount, a short expiry and an Ed25519 signature verifiable offline. Pre-payment AML/KYT clearance and counterparty due-diligence for autonomous agents. Input: to (counterparty wallet) + amount_usd (+ optional depth).

### GET /agent/visibility-audit
- what: Audit how discoverable an x402 agent/seller is across the agent registries (CDP Bazaar, 402index): keyword-RELEVANCE rank per category (not raw settled-volume rank), a metadata-quality score of your advertised endpoints (schema, output.example, tags, llm_usage_prompt), settle activity, a top-3 benchmark, prioritized fixes, and a DELTA vs a signed snapshot you carry back. Why am I not being found and how do I climb? GEO/AEO discovery audit for x402 sellers, Ed25519-signed.
- price: $1.00
- input:
  - seller: Seller to audit: wallet (0x + 40 hex) or origin URL/domain, e.g. 'api.example.com' or '0x1D1B...620f'
  - snapshot: Optional: the signed_snapshot JSON from a previous audit, to compute a dated delta
- source: Seller /.well-known/x402.json discovery + CDP Bazaar discovery/search + 402index + internal seller-trust + Claude Haiku (summary) + Ed25519 signed snapshot
- health: GET /agent/visibility-audit/health
- llm_usage_prompt: Discoverability audit of an x402 agent/seller across the agent registries (CDP Bazaar, 402index): keyword-relevance rank per category (not raw settled-volume rank), metadata-quality score of advertised endpoints (description density, typed input schema, output.example, tags, llm_usage_prompt), on-chain settle activity, top-3 competitor benchmark, impact-prioritized fixes, and a signed delta vs a previous snapshot you carry back. Why am I not being found and how do I climb? Discovery/GEO/AEO audit for x402 sellers. Input: seller (wallet or origin domain) (+ optional prior signed snapshot).

### GET /agent/rank-check
- what: Quick check of where an x402 seller ranks RIGHT NOW by keyword-relevance in the CDP Bazaar discovery (not the raw settled-volume rank the free explorers show): best rank + per-category-keyword rank in one cheap call, plus a pointer to the full /agent/visibility-audit when the rank slips. The frequent pulse for monitoring your x402 discoverability. Where do I rank now? Am I being out-ranked on my category keywords?
- price: $0.10
- input:
  - seller: Seller to check: wallet (0x + 40 hex) or origin URL/domain, e.g. 'api.example.com'
- source: CDP Bazaar discovery/search (keyword-relevance rank per category) + Ed25519 receipt
- health: GET /agent/rank-check/health
- llm_usage_prompt: Quick keyword-relevance rank of an x402 seller in CDP Bazaar discovery (not the raw settled-volume rank the free explorers show): best rank + per-category-keyword rank in one cheap call, plus a pointer to the full visibility-audit when the rank slips. The frequent pulse for monitoring x402 discoverability. Input: seller (wallet or origin domain).

### GET /agent/passport
- what: Measured-vs-declared reputation dossier for an AI agent, Sybil-corrected. Give an ERC-8004 agent id or wallet: reads the on-chain ERC-8004 Identity + Reputation registries on Base, computes the DECLARED reputation (reviews/reviewers/values) then the MEASURED, Sybil-corrected reputation (owner agent-farming, footprint-less reviewers, review concentration, unbacked-by-settlement), returns the DELTA + a signed offline-verifiable verdict. Real or inflated? Audit-grade due-diligence.
- price: $1.00
- input:
  - agent_id: ERC-8004 agent identity tokenId (IdentityRegistry). Provide this OR wallet.
  - wallet: Agent wallet (0x + 40 hex); resolved to its ERC-8004 agentId. Provide this OR agent_id.
- source: ERC-8004 Identity + Reputation registries on Base (eth_call) + Blockscout + internal seller-trust + Claude Haiku (narrative) + Ed25519 receipt
- health: GET /agent/passport/health
- llm_usage_prompt: Measured-vs-declared ERC-8004 agent reputation dossier, Sybil-corrected: reads the on-chain ERC-8004 Identity + Reputation registries on Base, computes the declared reputation (review count/reviewers/values) then the measured, Sybil-corrected reputation (owner agent-farming, footprint-less reviewers, review concentration, value uniformity, placeholder tags, unbacked-by-settlement), and returns the delta + a signed offline-verifiable verdict. Is this agent's reputation real or inflated? Audit-grade counterparty due-diligence. Input: agent_id (ERC-8004 tokenId) OR wallet.