Batch static accessibility (WCAG 2.2) check of many UI files at once: a list of objects each with a filename, optional framework and markup. Returns a per-file pass, caution or block verdict plus a rollup with the worst verdict and totals by severity. Accessibility indicators, not a legal compliance guarantee.

Static accessibility (WCAG 2.2) check of a UI markup fragment for JSX, TSX, Vue, Angular or HTML. Detects missing image alt, unlabeled form controls, buttons and links with no accessible name, skipped heading levels, invalid ARIA, non-interactive click handlers, duplicate ids, positive tabindex and low static colour contrast. Returns a pass, caution or block verdict with per-finding WCAG criterion, element and line. Accessibility indicators, not a legal compliance guarantee.

Resolve a financial professional or firm name to ranked CRD candidates for disambiguation before an authoritative verify call. Each candidate carries its crd, name, current firm, broker and adviser scope, and a disclosure flag. Accepts a name, or a firm.

AI-search visibility (AEO) audit of a website for answer engines and agents. Fetches the target's own public surface and computes a composite visibility score 0-100 with a rating (poor, fair, good, excellent) across six axes: answerability, structured data, AI-crawler access, crawlability, llms.txt readiness and MCP well-known readiness, plus a prioritized list of fixes. Deterministic, no LLM. Visibility indicators, not a guarantee.

robots.txt AI-crawler access matrix: for the 2026 cohort of AI and answer-engine crawlers (GPTBot, ChatGPT-User, OAI-SearchBot, ClaudeBot, Claude-User, PerplexityBot, Google-Extended and more) reports allowed, blocked or unlisted per bot, whether the default rule blocks everything, declared sitemaps and a score. One network operation.

Focused check of a site's /llms.txt and /llms-full.txt: presence, size and llmstxt.org structural validity (an H1 title, a blockquote summary, sections of markdown links). Returns a readiness score plus fixes. Cheap sub-check of the full AEO audit.

US crop production, yield, acreage and grain stocks from USDA NASS QuickStats by commodity, with an optional state and year. Returns a normalized cited series plus the year-over-year change.

Fused agricultural signal for a commodity: the USDA Market News spot price is combined with the NASS production and stocks series into computed signals - spot change, price percentile over available history, and production and stocks year-over-year context.

Title and airworthiness due-diligence verdict for a US-registered aircraft from the public FAA Aircraft Registry. Give a tail number or serial and get a 0-100 title risk score, a low, medium, high or critical rating, a clear_to_transact flag, per-dimension signals, reasons and the cited registry fields. Deterministic, no LLM. A diligence signal, not a certified title search.

List the US-registered aircraft held by an owner name in the public FAA Aircraft Registry (fleet lookup). Give an owner or company name and get the matching aircraft, each with its tail number, serial and make and model. Registry metadata only. A diligence signal, not a certified title search.

Look up a US-registered aircraft in the public FAA Aircraft Registry by tail number (N-number) or serial and get a normalized record: registration status, registered owner, make, model, year, serial, engine, airworthiness date, type certificate and Mode-S code. Cheap first step before aircraft/diligence.

Cited answer (RAG): a natural-language question is answered by searching the public web, reading the top pages and synthesizing a grounded answer with a list of source URLs. Answer cites only the retrieved sources (anti-hallucination).

Verify an attestation produced by /notarize (either the EIP-712 or the EAS off-chain form) — pure compute, no network and no keys. Recovers the signer and reports whether it matches this service's attestor, whether the EAS UID recomputes, and (if you supply the original body as payload_base64) whether its sha256 matches the attested hash.

Reverse BIC / SWIFT code lookup: given a BIC, return the resolved bank name(s), the country and the domestic bank codes. Offline open-registry data (schwifty, MIT); no network or keys. A malformed BIC returns valid false with an error_type, never a 500.

Cheap structural IBAN validation: ISO 13616 checksum + per-country structure/length. Returns valid plus, on rejection, a granular error_type (invalid_checksum, invalid_length, invalid_structure, invalid_country), the country and the format spec. Offline (schwifty, MIT); no network or keys. A cheap-first entry before the fuller /bank/verify fusion route.

Bank account / IBAN verification in one call: ISO 13616 checksum + per-country structure validation, BIC / bank name / branch resolution, country details, SEPA-zone flag, and a deterministic 0-100 fraud/KYC risk score with a band, reasons and weighted signals. Offline open-registry data (schwifty, MIT). Structural + resolution + risk indicators only; NOT proof the account exists or who owns it, and no name matching. Input not stored.

History of US bank failures and assisted acquisitions from the public FDIC failed-bank list. Filter by bank name, FDIC cert number, 2-letter state or year and get matching records with the failure date, resolution type, city and state and the deposits, assets and estimated cost at failure. Covers the 2023 failures including Silicon Valley, Signature and First Republic.

Search US FDIC-insured banks by name and get ranked candidate institutions, each with its FDIC certificate number, active status, charter class, city and state and headline assets, deposits and return on assets. Cheap first step before bankhealth/verdict.

Financial-health verdict for a US FDIC-insured bank from public call-report data. Give a bank name or FDIC cert number and get a 0-100 health score (HIGHER is HEALTHIER), an A-F letter grade, derived proxy metrics (capital ratios, Texas-ratio proxy, uninsured-deposit exposure, ROA trend) and per-dimension signals, grounded to the reporting date. Deterministic proxies, no LLM, not CAMELS. Not financial advice.

Look up the known breaches an email address appears in, returning breach METADATA only (name, date, data classes) and never passwords or leaked personal data. Uses the keyed Have I Been Pwned account API; when no API key is configured the route returns available false and the keyless password and domain routes still work. Breach indicators, not a guarantee.

Composite account-takeover risk score (0-100) fusing a privacy-preserving password exposure check with a domain's known breach history. The password is hashed locally (k-anonymity, never transmitted). Returns ato_risk_score, a category, per-signal reasons and the raw signals so the agent can re-rank. Risk indicators, not a guarantee.

List the known public data breaches that affected a domain, from the Have I Been Pwned breach catalogue. Returns normalized breach metadata (name, date, account count, data classes, stealer-log and verified flags) plus a summary. This is which known breaches touched the domain, not whether a specific account is compromised. Breach indicators, not a guarantee.

Check whether a password has appeared in known data breaches, using k-anonymity: the password is hashed locally and only a 5-character hash prefix is sent upstream, so the plaintext is never transmitted, logged or stored. Accepts a password, a SHA-1 hash or an NTLM hash. Returns compromised, breach_count and the hash prefix. Credential-exposure indicators, not a guarantee.

Bridge exploit track record: for one bridge, its hack history (incidents, total USD lost, funds returned, worst technique); with no bridge, a ranked leaderboard of every bridge hack by USD lost. Sourced from DefiLlama's hack dataset, joined and normalized by us.

Corridor safety verdict: given a source chain, destination chain and optional asset/amount, returns every bridge covering that corridor, each with its deterministic risk score and safe/caution/avoid tier, sorted safe-first, plus the recommended bridge and the reasons. One call to pick the safest bridge for a hop.

Cross-chain bridge safety screen: given a bridge (Stargate, Across, Wormhole, Arbitrum, CCTP, ...), returns a deterministic 0-100 risk score and a safe/caution/avoid verdict fusing its trust model, TVL depth, exploit history and governance/upgrade risk, with the reasons. One call instead of reconciling L2Beat, DefiLlama TVL and hack post-mortems yourself.

Bridge TVL and per-chain liquidity concentration: total value locked plus the breakdown across chains and the largest single chain's share, from DefiLlama. A liquidity-depth and concentration signal for routing decisions.

Balance and unspent outputs for a Bitcoin address. Returns confirmed and mempool balance, funded and spent totals, transaction count and a capped UTXO set for a legacy, p2sh or bech32 address. Informational, not advice.

Bitcoin block metadata by height or hash (or the chain tip when neither is given), fused with the current difficulty-adjustment context, a computed halving countdown and a network hashrate snapshot. Informational, not advice.

Recommended Bitcoin transaction fee-rates in sat per vByte across five confirmation horizons (fastest, half hour, hour, economy, minimum) plus a live mempool congestion snapshot with an estimated block backlog. Informational, not advice.

Live Bitcoin mempool snapshot: unconfirmed transaction count, total virtual size and total fees, a fee-rate histogram broken into buckets, and an estimated block backlog. Informational, not advice.

Decoded Bitcoin transaction by txid: inputs and outputs, fee, size, weight and virtual size, replace-by-fee flag, and confirmation status with a confirmation count. Informational, not advice.

Forward-looking FDA regulatory catalyst calendar for one biotech issuer, extracted from primary SEC 8-K and 6-K filings. Give a ticker or CIK and get upcoming PDUFA target action dates, advisory committee meetings, Complete Response Letters, priority reviews and PDUFA extensions, each date grounded in a quote from the filing and cited to its accession and EDGAR URL.

Extract every cited FDA regulatory-decision date from ONE SEC filing. Give an accession with its CIK, or a ticker or CIK to use the most recent matching 8-K or 6-K, and get all PDUFA, advisory committee, CRL, priority review and PDUFA extension dates in that document, each grounded in a quote. Cheap single-document pull.

Market-wide upcoming FDA regulatory catalysts across recent SEC 8-K and 6-K filings in a forward window. Get PDUFA dates, advisory committee meetings and other regulatory events for many biotech issuers at once, each grounded in a filing quote and cited to its accession and EDGAR URL. Bounded fan-out.

Cheap raw normalized reserve breakdown of one centralized exchange: total on-chain reserves (currentTvl), clean-of-own-token reserves (cleanAssetsTvl), the own-token symbol / USD value / concentration, clean-asset ratio, spot & derivative volumes, leverage, 24h/1w/1m netflow and the public reserves URL, with citations. The cheap first grain before the fuller /cex/solvency verdict. Keyless DefiLlama source; reserves are a proxy, not proof of solvency.

Market-wide ranked solvency screen of all centralized exchanges from DefiLlama CEX Transparency: each row has solvency_score, rating, clean-asset ratio, own-token concentration, 1-week netflow percent and risk flags. Filter by min_score / max_own_token_pct / min_tvl; sort by score, tvl or outflow. Deterministic; reserves are a proxy, not proof of solvency.

Centralized-exchange (CEX) solvency / proof-of-reserves counterparty-risk verdict in one call: given an exchange name, returns reserve totals, own-token concentration (the FTX/FTT pattern), clean-asset ratio, netflow stress, leverage and a deterministic 0-100 solvency_score with a rating, risk flags and a peer rank. Fused from the keyless DefiLlama CEX Transparency reserve aggregate. Reserves are a PROXY: customer liabilities and PoR-attestation quality are not visible.

Is an address a contract or an EOA on Base, Ethereum, Arbitrum, Optimism, Polygon or BNB (default Base)? Returns is_contract, deployed code size in bytes, and the keccak256 code hash (eth_getCode).

Static application-security scan of a single code blob for CWE Top-25 logic bugs (SQL injection, XSS, command injection, SSRF, path traversal, insecure deserialization, weak crypto and more). Pass the code and its language; returns a verdict with per-finding CWE, severity and line. Static indicators, not a guarantee.

Static application-security scan of source code or a git-diff for CWE Top-25 logic bugs: SQL injection, XSS, command injection, code and template injection, SSRF, path traversal, insecure deserialization, weak crypto, insecure randomness, open redirect and XXE across Python, JavaScript, TypeScript, Java and Go. Returns a go/no-go verdict with per-finding CWE, severity, file and line. Static indicators, not a guarantee.

Normalized US House congressional stock trades from STOCK Act Periodic Transaction Reports. Filter by ticker or member and an optional date window and get each disclosed trade: buy, sell or exchange, the amount range, ticker, asset description, transaction and disclosure dates, disclosure lag, and owner, each record cited to its filing PDF.

Pure-compute data format conversion (no network): JSON<->CSV, JSON<->YAML, JSON normalize/minify, Markdown->HTML (sanitized), HTML->Markdown/text, base64/hex, TOML->JSON. Operates only on the supplied data.

Normalized crime profile for one FBI reporting agency by ORI code. Give an ORI and an optional offense and get the agency's multi-year annualized offense rates per 100,000 people, the latest complete reporting year, the trend and the data as-of date. The raw grain behind crime/location, without geocoding.

National or state baseline crime rates for an offense and window from public FBI UCR and NIBRS data. Give an offense and an optional 2-letter state and get the annualized rate per 100,000 people by year, for anchoring a local score. Deterministic, no AI.

Recent incident density for a supported US city from its open-data portal. Give a city slug and an optional look-back window in days and get aggregated counts of reported incidents by offense category, a fresher signal than the lagged federal data. Counts by type, not a per-incident dump.

Crime-risk verdict for a US location from public FBI UCR and NIBRS data. Give an address, a latitude and longitude, an agency ORI code, or a state, and get a 0-100 crime risk score, a low, medium, high or critical rating, a violent and property breakdown, a multi-year trend and a comparison with the national baseline, citing the reporting agency and the data as-of date. Deterministic, no AI. Aggregate risk indicators, not a consumer report.

ABI-encode a function call (selector + args) or encode/decode arbitrary ABI types — build and parse EVM calldata correctly.

Convert token units (wei/gwei/ether or arbitrary decimals), between hex and int, and compute/validate an EIP-55 checksummed address.

Compute keccak256/sha256, a 4-byte function selector, an event topic0, or an ENS namehash/labelhash from a string or hex input.