Generate a 1D barcode as PNG or SVG: EAN-13, EAN-8, UPC-A, Code 128, Code 39 or ITF-14. Validates the structure and check digit for each symbology — EAN/UPC/ITF compute a missing check digit or reject a wrong one. Options: format (png/svg), scale (module px 1-10), text (human-readable line, default on). Returns the raw image. Generation only — reading a barcode from an image is out of scope.

Run up to 10 internal calls in one payment: POST {calls:[{path,input}]} and pay the SUM of the internal prices (≤ $0.05 total), executed in parallel, results returned in order. Batchable endpoints are the local, deterministic ones (validators, format transforms, time/locale) — no browser, network, composite or dynamic endpoints. Each result carries {path, status, body}; a bad request (unknown path, >10 calls, over cap) is a 400, not charged.

Business-day math with real holiday calendars for FR, DE, GB, US, ES, IT, NL, BE (2025-2027, official sources — same datasets as /v1/holidays). Count working/weekend/holiday days in a date range (?from&to&country) or shift a date by N business days (?from&add=10&country). Weekend configurable (?weekend=fri,sat), optional ?region= for Länder/UK divisions. Holidays that suppress a working day are listed by name.

Verify C2PA / Content Credentials (provenance & tamper detection) in an image: send the image bytes or a JSON {url}. Returns Trusted/Valid/Invalid/None with per-manifest signature, trust-chain (issuer) and asset-hash checks plus C2PA actions. 'Trusted' only if the signer chains to the C2PA trust list, otherwise 'Valid'. Relevant to EU AI Act Art. 50. Verification only (embedded manifests) — not generation.

Structural card-number check before hitting a PSP: Luhn checksum, network detection from public IIN prefixes (visa, mastercard, amex, discover, jcb, diners, unionpay) and per-network length validation. POST body on purpose: the number never appears in a URL, so it stays out of request logs and x402 payment metadata — never stored, logged or echoed back. No BIN/issuer lookup (BIN databases are licensed): well-formed does not mean the card exists.

Split a long text into token-sized chunks for RAG ingestion. Chunks are measured in REAL tokens (tiktoken o200k_base/cl100k_base), honour a token overlap, and cut on sentence, paragraph or word boundaries (strategy). Returns each chunk text, its exact token count and char offsets. The natural composite of token-count; counts are OpenAI-style estimates (no Claude tokenizer exists). Provided text only; nothing is fetched. Over 1 MB is truncated.

Colour utilities: ?op=convert (default) turns any hex/rgb()/hsl()/CSS-name into hex, rgb, hsl, hsv, alpha, relative luminance and exact CSS name; ?op=contrast with ?fg=&bg= returns the WCAG 2.x contrast ratio and AA/AAA pass/fail for normal & large text plus UI components (accessibility / EAA); ?op=palette derives complementary, analogous, triadic, tints and shades. Send hex without '#'.

Country reference card by ISO code (alpha-2/alpha-3/numeric) or English name: ISO 3166 codes, official name, current ISO 4217 currency (Bulgaria = EUR since 2026-01-01), dialling prefix, IANA ccTLDs (incl. IDN), languages, EU/EEA/Schengen/euro-area membership and the EU VAT prefix (Greece = EL). 56 countries: EU-27, EEA/EFTA, GB, US + 24 majors. Compiled from ISO/IANA/ITU/europa.eu sources, dated `asOf`.

Next occurrences of a cron expression in any IANA time zone, DST-aware via the runtime's ICU tzdb (times skipped by DST shift forward by the gap; ambiguous times fire once). Dialect (croner): 5-7 fields with optional seconds and year, ranges, steps, lists, JAN-DEC/SUN-SAT names, L, W, #, ? and @daily-style aliases. Each occurrence is returned as local ISO with offset plus UTC. Bad expression → 400 naming the offending field.

Row-level diff of two CSVs (a→b) — reconcile two exports. With `key` (a column name, or a column index when header:false), rows pair by key → added / removed / changed (per-field old→new). Without a key, a whole-row multiset diff gives added/removed only. Delimiter auto-detected (, ; tab |); tolerant RFC-4180 parsing. First row is the header unless header:false. Max 5 MB per input; reported buckets cap at 10000.

Profile a CSV before processing it: send the bytes directly or a JSON body with a public {url}. Returns per-column name, inferred type (string/int/float/bool/date/mixed), null and unique counts, min/max and samples, plus row count, detected delimiter and data-quality issues (malformed rows, inconsistent widths, duplicate headers). Query: delimiter (, ; tab |), header=false if row one is data. Max 10 MB.

Convert CSV to JSON with a strict RFC 4180 parser: quoted fields, doubled quotes, delimiters and line breaks inside quotes, mixed CRLF/LF, BOM. Auto-detects comma, semicolon, tab or pipe (or set delimiter). First row becomes object keys; headers:false yields string arrays. Ragged rows and unterminated quotes are repaired and reported in errors[], never a crash. All values stay strings. Send raw CSV text or JSON {csv}. Max 5 MB.

De-duplicate a list of records or strings. Exact: group by a `key` field (dot-path) or, by default, the whole item (key-order-insensitive). Fuzzy (`fuzzy`: a field name, or true for scalar strings): group by normalized Levenshtein similarity ≥ `threshold` (default 0.85) — e.g. 'Jean Dupont' ≈ 'jean dupont' but 'Jean Martin' stays separate. Returns unique[] and duplicates[] (kept + dropped indices). Fuzzy ≤ 1000 items.

Compute the diff between two texts (a → b) — no git needed. Returns a standard unified patch (line-oriented in both modes, applies with git apply or patch), added/removed counts, and in word mode a token-level change list for inline highlighting (capped at 2000 entries). CRLF line endings and missing trailing newlines are preserved byte-exactly. Caps: 1 MB UTF-8 per input; pathologically divergent inputs return 422 diff_too_complex.

Resolve DNS records for a domain over DNS-over-HTTPS: A, AAAA, MX, TXT, NS and CNAME, one or several types per call (type=A,MX). Records come back normalised — MX priority split out, trailing dots stripped, TXT quoting removed — each with its TTL. Cloudflare DoH with automatic Google DoH fallback for resilience; NXDOMAIN is reported explicitly instead of as an error.

Extract text from a .docx (Word) document: send the docx bytes directly, or a JSON body with a public {url}. Returns the plain text, a best-effort markdown rendering (headings, bold/italic and lists; tables are flattened to paragraphs), and metadata (title, author). Text layer only — no OCR of embedded images. Max 10 MB.

Domain availability and expiry intelligence via RDAP, the WHOIS successor. Queries the authoritative registry for any RDAP-enabled TLD: free domains return registered:false, taken ones return registrar, creation and expiry dates, nameservers and EPP status codes. TLDs without RDAP coverage return supported:false. Built for agents that buy, monitor or vet domains.

Support 402utils — or use it as a parametric x402 payment-test endpoint: you set the price with ?amount= (USD, clamped to $0.001–$0.05) and the 402 demands exactly that, so you can assert a client pays the precise amount and the on-chain USDC tx matches. After payment returns {thanks, amount, message}. The amount is a pure function of ?amount, identical on the unpaid probe and the paid retry.

TARGET2/T2 settlement calendar — the SEPA banking calendar, to predict when a euro transfer settles. TARGET is closed on weekends + six fixed days (New Year, Good Friday, Easter Monday, 1 May, 25 & 26 Dec). ?year= lists the closing days; ?date= says whether it settles (and the next open day); ?date=&addBusinessDays=N returns the value date N settlement days away. The SEPA counterpart of /v1/business-days.

Echoes back the JSON body you send, with a server timestamp. Useful for verifying your x402 client integration end to end.

Validate an email address before you accept a signup or send to it: RFC 5322 syntax, whether the domain resolves and its MX records, whether it is a disposable/throwaway domain (community blocklist), and whether it is a role account (info@, support@…). DNS is checked over DoH. No SMTP probing (intrusive and spammy), so a true 'domainExists' does not guarantee the mailbox exists.

Text → vector embeddings via Cloudflare Workers AI. Send {text} or {texts[]} (≤20, each ≤4000 chars); returns one float vector per input. Default model @cf/baai/bge-m3 (1024-dim, multilingual). The model is named in the response — embeddings only compare within the same model. Completes the RAG chain: readability → chunk → embed, no OpenAI account. Cosine-compare the vectors yourself. Unavailable model/binding → 503, not billed.

Validate an EU EORI number (the customs operator id for any EU import/export) against the Commission's official EOS service. Format is checked locally first, so junk never hits the network. Valid numbers return the registered name/address when the operator consented. Never a false 'invalid': if the service is down you get a 503, not charged. Complements vat-check in the EU compliance toolkit. ?attest=1 → Ed25519-signed.

Generate an EPC QR code (SEPA Credit Transfer / « Girocode », EPC069-12) — the QR a banking app scans to pre-fill a transfer, the payment companion to an invoice. Body { name, iban, amount?, bic?, purpose?, remittance? | reference? (RF) }. Returns PNG (default), SVG, or the raw EPC text. IBAN + RF are validated; strict EPC encoding (order, lengths, ≤331 bytes, level M). EUR only. Not a payment initiation — just the encoded data.

Screen a name against the EU consolidated financial sanctions list (embedded official snapshot; persons + entities). Fuzzy matching returns each hit with a score, sanction programme, listing date and the matched name. POST body {name, type?, threshold?} — POST so the name never enters logs. A SCREENING AID against the EU list only (not OFAC/UN/UK); a match is not a compliance determination — verify with the official source before acting.

Extract emails, http(s) URLs, international phone numbers, @mentions and #hashtags from provided text OR html — normalized and de-duplicated. From html, also reads href/src attributes and mailto: links. Emails are syntax-validated; phones use Google libphonenumber (international format). Pattern-based extraction, NOT ML named-entity recognition. What a crawl agent pulls from every page. Processes provided content only; nothing is fetched.

Read a Factur-X / ZUGFeRD invoice: send PDF bytes or a JSON { url }. Extracts the embedded EN 16931 CII into normalised data — seller, buyer, lines, VAT, totals, payment, dates, references — with the detected profile. The read side of e-invoicing (the FR mandate starts with reception). No embedded CII → { profile:"none", warning:"no_embedded_cii" }. ?attest=1 → Ed25519-signed.

Generate a Factur-X invoice (PDF/A-3 + embedded EN 16931 CII XML) from structured data: send { invoice, profile }, get the PDF bytes (or JSON with base64 + conformity report via Accept: application/json). Targets EN 16931 and the French PA « Flux 2 » rules (payment mentions, business process, electronic addresses) ahead of the 2026-09 mandate. Technical artifact, not tax/legal advice; you remain responsible for data accuracy and PA transmission.

Validate a Factur-X / ZUGFeRD invoice: send PDF bytes, JSON { url } or { pdfBase64 }, or raw CII { xml }. Returns a rule-based report — PDF/A-3 structure, EN 16931 CII anchors, arithmetic rules (BR-CO) and French PA « Flux 2 » checks (BR-FR) — with the detected profile. Structural + heuristic, not a certified conformance service. ?attest=1 → Ed25519-signed audit trail.

Resolve a website's best icon: parses <head> for link rel=icon / apple-touch-icon / mask-icon, reads a linked web app manifest's icons, and adds the /favicon.ico fallback, then ranks them (scalable/large first). Returns the best iconUrl plus all ranked candidates with rel/sizes/type/source. The URL is SSRF-checked; fetches are capped. For agents building directories or UIs.

Daily euro FX reference rates from the European Central Bank. ?base (default EUR), ?quote for a single rate, ?date=YYYY-MM-DD for history (last 90 ECB business days; older dates 404). Non-EUR bases are cross-rates derived via EUR (quote-per-EUR ÷ base-per-EUR). Weekend/holiday dates resolve to the last published ECB day — the date field says which. Published ~16:00 CET on ECB business days; informational rates, not for transaction pricing.

Compute cryptographic hashes / checksums of a string or of raw bytes (base64). Returns lowercase hex for each requested algorithm: sha256 (default), sha512, sha1, md5. SHA via native Web Crypto; md5 and sha1 are offered for legacy interop only and flagged — not for security. POST so the input never lands in a URL/log. Verification/checksums only, not a hash cracker.

Public and bank holidays for FR, DE, GB, US, ES, IT, NL, BE — years 2025-2027, each dataset verified against the country's official source (Etalab, Land statutes, gov.uk feed, OPM, BOE, normattiva, rijksoverheid, SPF Emploi). Movable feasts are computed (Gregorian computus), substitute days follow each country's official rule. Optional ?region= for sub-national holidays (German Länder, UK divisions, Alsace-Moselle).

Extract every table from HTML you provide into clean JSON — no fetching, send the markup itself as {html} or a raw text/html body. Returns per table: optional caption, headers[] (from thead or an all-th first row) and rectangular rows[][]. colspan/rowspan cells are grid-expanded with the value duplicated so rows align with the visual layout; nested tables are flattened into the outer cell's text. Caps: 5 MB HTML, 200 tables, 200k cells.

Convert raw HTML into clean GitHub-Flavored Markdown: ATX headings, fenced code, pipe tables, strikethrough, task lists; script/style/iframe text is always stripped. mode=article isolates the main article with Mozilla Readability (returns title, byline, excerpt; drops nav/footer boilerplate; resolves relative links), else converts the whole document. Converts provided HTML only — never fetches URLs. Max 5 MB.

Validate an IBAN structurally before you pay or store it: ISO 7064 mod-97 check digits plus the official per-country length and BBAN layout (~90 countries), returning the bank/branch/account breakdown for FR, DE, ES, IT, BE and NL. Structural validation only — no BIC directory lookup (SWIFT's BIC registry is proprietary) and no proof the account exists. Local, instant. ?attest=1 → Ed25519-signed.

Build a strictly RFC 5545-conformant .ics calendar file from a list of events: correct TEXT escaping, 75-octet line folding, generated UIDs, and unambiguous UTC timestamps (a per-event IANA timezone interprets naive wall-clock times). Supports all-day events, location, description and a passthrough RRULE. The fiddly conformance an agent shouldn't hand-roll.

Resize and convert an image (PNG, JPEG or WebP, up to 10 MB). Send the image bytes as the body and set options in the query: width, height, fit (contain or cover), format (webp/jpeg/png) and quality (JPEG). Returns the converted image bytes. WebP has no quality knob and AVIF is not supported (out of scope).

Pre-flight a SEPA invoice's identifiers in one paid call: validate the customer VAT (VIES, with company name), the IBAN (ISO 7064 mod-97 + BBAN layout) and an ISO 11649 RF creditor reference, run together with a per-field status. summary.allProvidedValid is true only if every provided identifier checked out. If all provided checks fail technically you get a 503 and no charge. ?attest=1 → Ed25519-signed.

Apply a jq-style filter to JSON — the JSON swiss-army knife. Supports identity, fields, index/slice/iterate, pipe |, comma, //, and/or/not, arithmetic, comparisons, array/object construction and builtins (select, map, keys, length, add, sort_by, group_by, unique, to_entries, has, range…). A SAFE pure-JS subset (no eval, bounded): no variables, def, reduce, regex or assignment. Returns the whole jq output stream as results[].

Convert a JSON array of objects into a clean RFC 4180 CSV: header row, CRLF records, correct quoting and quote-doubling. Nested objects flatten to dot-notation columns (a.b.c), arrays are JSON-encoded into their cell, and you can select and order columns explicitly. Delimiter: comma, semicolon, tab or pipe. Caps: 100,000 rows, 5 MB body, 20 MB CSV out. Returns text/csv ready for Excel, Google Sheets or a data pipeline.

Reshape JSON declaratively — predictable, not a language. `mapping` mirrors the output shape; its leaf strings are dot-paths into `data` (a.b.0.c; negative index ok; 'contacts.*.email' projects over an array; '=text' escapes a literal). Objects/arrays nest; number/boolean/null are literals. Missing paths → null. Field rename/flatten/projection to fit a payload to a target schema. Distinct from /v1/jq (declarative vs query language).

Validate a JSON value against a JSON Schema (draft 2020-12 default; 2019-09/7/4 via $schema). Send {data, schema}; returns {valid, errors:[{path, message, keyword}]} with a JSON Pointer path per failure. For agents checking structured/tool-call output before use. Runs on an eval-less validator (Ajv can't run on Workers). Remote $ref is refused (self-contained only); unsafe regex patterns are rejected before compilation.

Extract values from a JSON document with a JSONPath expression or an RFC 6901 JSON Pointer, without returning the whole document. Send {data, path}; get {matches, count}. Path starting with $ = JSONPath ($.items[*].price, $..author, [0], [1:5], [*]); starting with / = JSON Pointer (/items/0/price). No-eval tree walk with bounded traversal; filter expressions [?(…)] are not supported by design.

Extract salient keywords/key-phrases from text via TextRank (word co-occurrence graph + PageRank), merging adjacent high-ranked words into phrases. Pure, deterministic, model-free — extractive, NOT ML/NER or an LLM (the point: reproducible, no inference cost). English + French stopwords removed. Returns keywords[] with normalized scores (top ≈ 1). For content tagging/indexing and pre-RAG. Over 100k tokens is truncated.

Detect the language of a text with franc (pure-JS trigram model, 61 languages + CJK/Indic scripts). Returns the ISO 639-1 code (639-3 when no 2-letter code exists), language name, a separation-margin confidence (0=ambiguous, 1=unambiguous), a reliable flag, and ranked alternatives. The pre-sort step of any multilingual pipeline. Short/ambiguous text returns reliable:false honestly. Processes provided text only; nothing is fetched.

Global KYB via the official GLEIF registry. ?lei=<20-char LEI> returns legal name, entity + registration status, jurisdiction, address, next renewal and direct parent; ?name=&country= searches by name. The ISO 17442 checksum is verified locally, so a typo is a 400 with no network call. GLEIF down → 503, unbilled. Public CC0 data. Part of the EU compliance toolkit; enriches supplier-check.

Verify a link before you cite or use it: HTTP status, the full redirect chain, the final URL, latency, and the final Content-Type/Content-Length. Redirects are followed manually (max 10) with SSRF validation on every hop; probes with HEAD by default and falls back to GET automatically when a server rejects HEAD. A dead or broken target (404, 410, 500…) is a valid result, not an error. Response bodies are never downloaded.

Fetch and parse a site's llms.txt (llmstxt.org) — the LLM-era robots.txt. Give any page URL; it resolves the origin's /llms.txt, and returns the H1 title, the blockquote summary, and every section's links ([title](url): notes). Also probes /llms-full.txt and reports its size. A site without one returns {found:false} (a 200, still useful). The fetch is SSRF-checked. Bonus signal in /v1/site-profile.

Live market price stats for a kind of x402 service, from the public Bazaar catalog. Pass ?category=vat or ?query=pdf extraction and get {sampleSize, priceUsd:{min,p25,median,max,p75}, networks, examples} in USD — is your price fair, what should you charge? Computed over a live sample of up to 1000 resources; <3 matches returns an honest 'insufficient market data'. Neutral: our own endpoints are never featured.