johtodata.com

Fetch one public web page and return clean Markdown plus title. For LLM context and RAG. Tables: HTML tables become GitHub-flavored Markdown with a | --- | header row. Nested tables are lifted out as labelled sub-tables under ### Parent row → column headings, with (see Parent → column) in the parent

Fetch one public web page and return emails, phone numbers and social profile links found on it.

POST documents or text chunks plus a token budget. Returns deduplicated, query-ranked, whitespace-compressed context that fits the budget, with source offsets and a token estimate. Deterministic, no LLM.

POST CSV text or JSON rows. Returns inferred schema plus null, missing, duplicate, type-mismatch and numeric outlier findings per column, and a 0-100 quality score. Never changes values.

POST a URL of an x402 paid endpoint, or a raw PAYMENT-REQUIRED header or 402 JSON body. Returns PASS/WARN/FAIL checks for x402 v1/v2 (header/version, CAIP-2 network, USDC asset, atomic amount, payTo, timeouts, EIP-712) plus optional live checks when url is set: facilitator /supported lists network+s

POST broken or LLM-truncated JSON (max 16 KB). Returns valid JSON plus an ordered list of fixes: code fences, comments, single quotes, bare keys, trailing or missing commas, Python literals and unclosed strings, arrays or objects. Valid input is returned unchanged.

POST two JSON Schemas or OpenAPI documents as old and new. Returns breaking and non-breaking changes with JSON pointers and reason codes: removed properties, new required fields, narrowed types, enums and bounds, removed paths, operations, parameters and responses.

POST text, HTML or a public URL. Returns a 0-100 prompt-injection risk score with rule hits and exact character spans: instruction overrides, forged role tags, system-prompt leaks, secret exfiltration, hidden HTML text, zero-width and Unicode tag characters. Deterministic, versioned rules.

POST text (max 64 KB). Detects and redacts emails, phone numbers, Luhn-valid card numbers, SSNs, IBANs (mod 97), IPv4/IPv6 addresses, API keys, JWTs, private keys and password assignments. Returns redacted text, per-type counts and findings with offsets. Nothing stored.

POST a YAML, JSON, TOML, .env, Dockerfile or GitHub Actions workflow (max 64 KB, kind auto-detected). Returns syntax errors and common mistakes with line numbers and rule IDs: tabs, duplicate keys, unpinned images and actions, root user, curl-pipe-shell, script injection, committed secrets. Optional

POST up to 50 identifiers. Validates IBAN (mod 97), EU/UK/CH/NO VAT format, EAN/UPC/ISBN check digits, card numbers (Luhn, brand), EVM addresses (EIP-55 checksum), Solana base58 addresses, email syntax and E.164 phone numbers. Type auto-detected when omitted. Returns normalized values.

POST HTML or a public URL. Strips scripts, styles, nav and footers, converts to Markdown and returns heading-aware chunks with heading paths, token estimates, stable hashes and configurable overlap. For embedding and RAG pipelines. Tables: HTML tables become GitHub-flavored Markdown with a | --- | h

Rank up to 200 caller-supplied documents with lexical BM25 and return top matches, scores and exact snippet offsets. No web search, embeddings or external fetches. Caller must have rights to supplied text.

Extract application/ld+json script blocks from caller-supplied HTML, preserving parsed fields and source offsets. Reports malformed blocks. No URL fetch, JavaScript execution or remote context resolution. Caller must have rights to supplied HTML.

Decode one caller-supplied EVM ERC-20 Transfer log into from, to and exact atomic value. Optional decimals formats a decimal string. No RPC, chain verification, token identity check or trading. Works with standard logs on EVM chains.

POST a public .docx URL or base64 DOCX (max 10 MB). Returns Markdown text plus image attachments. Caller is responsible for the target site's terms. robots.txt is respected. Private hosts and redirects to them are blocked.

Prepaid API credit pack. Pay $0.99 USDC via x402 and receive a one-time API key with 200 calls on Starter routes. Stack packs by buying again. For monthly Starter, Growth or Scale plans email [email protected].

Search official CPSC product recalls, products and hazards for cross-border sourcing. Chinese field labels. Not certification or safety clearance.

Search official FDA food, drug or device enforcement recalls. Product, classification, status and dates. No patient records or medical advice.

Find USITC HTS codes, general/special rates and footnotes. Chinese field labels. Not total China-origin duties or customs advice.

Search Federal Register trade and regulatory publication metadata. Dates, agencies and official PDF links. Not legal advice.

Calculate landed cost and margin from caller-supplied costs, duty/VAT/fee assumptions. Exact decimal arithmetic. No live rate lookup or legal advice.

Validate GTIN-8/12/13/14 check digits for product catalogs, preserving leading zeros. No registry, ownership or authenticity verification.

Read attributed, dated ECB EUR-base currency reference rates as unchanged strings. Information only, not transaction quotes. No payment processing.

Compare two caller-supplied JSON or CSV product snapshots by SKU. Added, removed, changed products and currency-aware price deltas. No scraping.

Calculate dimensional and billable kg from caller-supplied centimetres, actual kg, divisor and optional rounding step. No carrier price lookup.

Normalize caller-supplied product titles using Unicode NFKC and whitespace cleanup. Chinese supported, originals preserved. No translation or truncation.

POST a company domain or ticker. Returns cited firmographics: name, legal name, industry, founded year, employees, HQ, country, tickers, SIC, latest annual report and revenue, LEI and company socials, each field with a source URL. Sources: company homepage, Wikidata, SEC EDGAR. No personal contacts.

POST one page URL on a site you own or are authorized to audit. Returns on-page SEO as JSON: title, meta description, canonical, noindex, robots.txt and sitemap status, headings, images missing alt, links, schema.org types, hreflang, Open Graph, coded issues and a 0-100 score.

Official USGS earthquakes near a point: magnitude, depth, felt reports and time. Bounded date/radius search, newest first. Public-domain source; not a safety forecast.

Active official US National Weather Service alerts at a US point: severity, urgency, onset and expiration, most severe first. Public-domain source, not an emergency warning service.

Look up organization LEI, legal name, jurisdiction and registration status from the official GLEIF API under CC0. No people, contact data or addresses. Name or exact LEI.

Official NHTSA safety recall campaigns by make, model and year: components, hazards, consequences and remedy. Model-level campaigns, not VIN repair status. Public-domain source.

Public DNS records over Google DoH. Up to four record types, DNS status, TTL and authenticated-data flag. Authorized domains only, no SMTP or mailbox verification. Email-shaped values are redacted.

MX, SPF presence and exact-domain DMARC policy diagnostics via DoH. Domain only, no mailbox addresses, SMTP or send. Not deliverability verification; SPF includes and organizational DMARC fallback not evaluated.

Exact crypto-address matching against the official OFAC SDN XML, up to 50 addresses. No identity data. EVM hex case normalized; other encodings case sensitive. Up to one-hour source cache. Non-match is not sanctions clearance.

Normalize and compare caller-supplied funding records from any labeled venue or chain with explicit units, intervals and direction. NOT a live feed. No exchange requests or unlicensed market-data resale. Simple annualization is not APY.

Recent SEC EDGAR filing metadata for public companies by CIK, optionally filtered by form. Dates, accession IDs and primary-document names only. No personal filers, contacts, addresses or filing text.

Official NIST NVD record by exact CVE ID, English description, CVSS metrics, weaknesses and CISA exploitation metadata. Not a vulnerability scan. NVD API attribution, no endorsement.

Recent official OpenFEMA v2 disaster declarations by US state, incident dates and program flags. Statistical research/reporting only, not individual eligibility or benefit decisions. FEMA attribution included.

Official NHTSA numerical crash and rollover rating facts by safety-rating vehicle ID. No VIN owner data, photos or copyrighted logos. Not a safety guarantee or repair-status check.

Official openFDA NDC directory facts by generic name: ingredients, dosage, routes, product identifiers and marketing category. Not full drug-label text or medical advice. NDC listing does not establish FDA approval.

Official openFDA 510(k) device-clearance factual metadata by device name. K numbers, dates, product code and decisions. No contacts, addresses or label text. Clearance is not premarket approval or medical advice.

One paid call for up to 20 company domains: public business contact pages, company emails/phones (free-webmail dropped), socials and light firmographics. Optional industry keyword match. Authorized collection only. Excluded from China MCP.

One paid call for a full company profile: homepage firmographics, Wikidata, DNS and email-domain diagnostics, on-page SEO signals and recent SEC filings when a US ticker is known. Company-level facts only.

One paid call for hiring signals across up to 10 Greenhouse, Lever or Ashby boards. Deduped job IDs, titles, locations and URLs. No LinkedIn scraping and no applicant PII.

One paid call for a product list: USITC HTS candidates, CPSC and FDA food recalls, optional landed-margin math and dated ECB FX (USD/CNY). Non-personal, China-market friendly. Not customs advice.

Diff an authorized page against a caller-supplied previous text or SHA-256. Returns changed flag, new hash and a short token diff. No screenshot.

Fetch up to 10 URLs from an authorized sitemap and return Markdown-ish text per page in one paid call. No JS rendering. Tables: this route does not emit GFM tables. It strips HTML tags to plain text, so nested tables are flattened and links lose their href.

Screenshot-free structured extract: title, meta description, price candidates, tables and JSON-LD from server HTML.