Aayat AI

Endpoints (146 live)

• POST /entities — Named-entity extraction: the organisations, people, places, products, events, dates, amounts of money and laws in a text or document (URL to a page or PDF), with how often each is mentioned. Every entity is checked by code to really appear in the source, so nothing is made up. Filter with types. About 12,000 characters read per call. (0.005 USDC on Base)
• POST /text/actions — Action items from meeting notes, an email thread, a transcript or a chat log: each task with its owner and due date (relative dates like 'next Friday' turned into YYYY-MM-DD when you pass today), plus the decisions made and open questions. Each item comes with the passage it was taken from, checked by code. Text or a URL; about 15,000 characters read. (0.01 USDC on Solana)
• POST /keywords — Key phrases and topics of a text, web page or PDF, for tagging, SEO and search indexes: up to 30 key phrases, most important first, each checked by code to really appear in the source (with how often), plus 3-5 broader topics. Any language. Reads about 12,000 characters. (0.003 USDC on Base)
• GET /perps/funding — Perpetual futures funding rates from Hyperliquid and dYdX: the current hourly rate, 8-hour and annualised equivalents, premium, mark and oracle price and open interest per market, plus up to 7 days of hourly funding history and averages for one coin. Pass ?coin=BTC&hours=24 (&venue=dydx), or no coin for the top markets by open interest. (0.003 USDC on Base)
• GET /vat/calc — Invoice-ready VAT maths: add VAT to net amounts or take it out of gross ones, per line and in total, rounded to the penny the way HMRC allows, with the amounts formatted in the currency and locale you choose. Pass ?amounts=19.99,5.00&rate=20&mode=net&currency=GBP. Any rate (UK 20/5/0, EU rates) and 150+ currencies. (0.001 USDC on Base)
• POST /receipt/scan — Photo of a receipt or invoice to clean JSON: shop or supplier, VAT/tax ID, date, currency, line items, subtotal, tax, total and payment details, with the maths checked by code (do the lines add up, does subtotal + tax = total). JPEG, PNG, GIF or WebP up to 5 MB from a public URL. For PDFs and text use invoice-parse. The text read from the image is included. (0.03 USDC on Solana)
• POST /marketplace/badge — For sellers listed on the Aayat AI marketplace: request the 'Verified by Aayat AI' badge for your approved listing. A person reviews the endpoint (it works as described, prices are clear, the payTo is yours); the badge shows once granted. One payment per listing. (1 USDC on Solana)
• POST /sentiment — Sentiment analysis for reviews, comments, support tickets and social replies: up to 50 texts per call, each labelled positive, neutral, negative or mixed with a score from -1 to 1 and its main emotion, plus totals. Name aspects (e.g. price, delivery, support) to get the sentiment towards each. Any language. About 12,000 characters in total per call. (0.005 USDC on Base)
• POST /invoice/parse — Parse an invoice or receipt into clean JSON: supplier, VAT/tax ID, invoice number, dates, currency, line items, subtotal, tax, total, amount due and bank details. Send a PDF, Word or web page URL, or the text (for a photo, run image-ocr first). Our code then checks the maths: do the lines add up, does subtotal + tax = total. Missing fields are null, never guessed. (0.02 USDC on Solana)
• POST /marketplace/feature — For sellers listed on the Aayat marketplace: feature your approved x402 API or MCP server for one more day per call ($1 a day). It shows in a clearly labelled 'Sponsored' slot next to matching router answers; the organic ranking itself never changes. Give your listing id. (1 USDC on Base)
• GET /papers/search — Search academic papers across Europe PMC (PubMed, PMC, preprints) and Crossref (all disciplines), merged by DOI: title, authors, year, journal, DOI, abstract, citation count and a free full-text link when open access. q is the topic or title words; sort is relevance, recent or cited; from_year and open_access=true narrow the results. ?q=retrieval augmented generation&limit=10 (0.003 USDC on Base)
• POST /moderate — Content moderation for AI agents and apps: is this text safe? Checks a user message (role=user) or an AI reply (role=assistant) against 14 hazard categories (violence, hate, self-harm, sexual content, child safety, weapons, privacy, elections...) with Meta's Llama Guard 3. Returns safe true/false and the categories hit. Up to 8,000 characters; failures are not charged. (0.005 USDC on Base)
• POST /pii/redact — Remove personal data from text before an agent stores it or sends it to another model: emails, phone numbers, card numbers, IBANs, NI numbers, SSNs, IP addresses, postcodes, people's names and street addresses become placeholders like [EMAIL_1] and [PERSON_2]. returnMap=true gives the placeholder list to restore them later. Up to 12,000 characters. Nothing is cached or kept. (0.005 USDC on Solana)
• POST /readability — Readability of English text or a web page/PDF: Flesch reading ease, Flesch-Kincaid grade, Gunning fog, SMOG, Coleman-Liau and ARI scores, a plain-words level, word and sentence counts, reading time, and what to fix: the longest sentences, long words and passive phrases. Pure code, no AI: the same text always scores the same. (0.002 USDC on Base)
• GET /perps/markets — Every perpetual futures market on Hyperliquid and dYdX side by side: which venues list each coin, price, open interest, 24h volume, max leverage and funding (hourly and APR) on each, plus the cross-venue funding spread and where longs pay least. Pass ?sort=fundingSpread&limit=20 for arbitrage candidates, or ?coin=SOL for one coin. (0.003 USDC on Base)
• GET /iban/check — Validate an IBAN before you pay: check digits (mod-97), the right length for its country, SEPA membership, and the bank code, branch and account parts (UK: sort code and account number). Pass ?iban=GB82WEST12345698765432 (spaces fine). Format check only: it can't confirm the account exists or its owner. (0.001 USDC on Base)
• POST /text/compare — What changed between two versions of a text: terms and conditions, a contract, a privacy policy, docs, a pricing page. Our code finds the exact sentences added and removed; AI then explains the changes that matter, most important first (e.g. 'notice period cut from 30 to 14 days'). Optional focus, e.g. 'fees and cancellation'. Up to 60,000 characters each. (0.01 USDC on Base)
• GET /http/security-headers — Security audit of a website's HTTP headers, graded A–F with a fix per finding: HTTPS and the http→https redirect, HSTS, Content-Security-Policy strength, clickjacking, nosniff, Referrer- and Permissions-Policy, cookie flags, version leaks, CORS. url is the site to check (https://example.com). Based on OWASP Secure Headers guidance. (0.003 USDC on Base)
• GET /weather — Weather now and forecast for any latitude/longitude on Earth: temperature, feels-like, wind, humidity, cloud, pressure, rain next hour, the next 12 hours, and daily min/max, rain and summary for up to 9 days. Pass ?lat=51.51&lon=-0.13&days=3. From MET Norway (the Norwegian national weather service), all times UTC. (0.002 USDC on Base)
• POST /language/detect — Detect the language of text: up to 20 texts per call, each with its ISO 639-1 code (en, fr, zh...), English name, confidence and writing script. Use it to route messages to the right translator, model or support team. Text with no letters returns language null. Reads the first 300 characters of each text. (0.002 USDC on Base)

+126 more endpoints.

MCP tools (15)

Aayat AI https://aayatai.com/mcp

• balance — Shows how many free calls are left today and the prepaid credit balance (with roughly how many paid calls it covers), plus any top-up still waiting for payment. Use when the user asks what's left, or after `topup` to see whether a payment has arrived; to add credit, use `topup` instead. Reads the key in the server URL if there is one, otherwise the free calls for this connection. Free to call. Pass `code` to redeem a credit code (ac_…) we gave back for an empty result: it's added to the key's balance (a new key is made if there isn't one).
• docs-answer — An answer written only from a library's current docs, with a code example and sources. ★ One of our best tools. Reads the project's current docs (llms.txt, docs pages or latest README) for any npm, PyPI, crates or Go library and answers from them alone, with a code example, the version it applies to and cited sources; says so when the docs don't cover it. library is the package name, question your question; ecosystem defaults to npm. Use for a direct answer to one how-to question about a library, from its current docs. Instead, use `docs-lib` ($0.005) for the docs to read yourself. Answers may be cached for up to 6 hours. Price: $0.02 USDC per call (x402 or prepaid credits); failed calls are free. Paid only (not in the free trial). Try docs-lib free first.
• docs-find — Find the AI-ready docs for any library or API: checks the project's docs site (from its npm, PyPI, crates or Go metadata) or any company domain (docs., developers., /docs) for llms.txt and llms-full.txt files, with titles, sizes and page counts, plus homepage and repository. Use to locate where a project's AI-ready docs (llms.txt) live. Instead, use `docs-lib` ($0.005) for the docs content itself. Answers may be cached for up to 24 hours. Price: $0.003 USDC per call (x402 or prepaid credits); failed calls are free. In the free trial.
• docs-lib — Stop your assistant using stale APIs: the latest docs for any npm, PyPI, crates or Go library, trimmed to your topic. ★ One of our best tools. Up-to-date docs for any npm, PyPI, crates or Go library: finds the project's own llms.txt and docs pages (or the latest release's README), keeps only the sections relevant to topic within a tokens budget, and returns them with sources and the current version. library is the package name; ecosystem defaults to npm. Use before writing code against a library, to read its current docs on a topic. Instead, use `docs-answer` ($0.02) for a direct answer to one question; `docs-find` ($0.003) for docs locations only. Answers may be cached for up to 6 hours. Price: $0.005 USDC per call (x402 or prepaid credits); failed calls are free. In the free trial.
• extract — Web page extractor for AI agents: fetch any public URL and get clean Markdown, the page title and all links as JSON, with scripts, styles and menus removed. url is the page address (https://...). Private or internal addresses are blocked; pages over 2 MB or slower than 10 seconds are rejected and you are not charged. Use to read one HTML web page as Markdown together with all its links. Instead, use `markdown` ($0.002) for PDFs and Office documents too. Answers may be cached for up to 10 minutes. Price: $0.005 USDC per call (x402 or prepaid credits); failed calls are free. In the free trial.
• find-best-tool — Describe a task and get the best pay-per-call APIs and MCP tools for it, ranked with prices and reliability. Find the best tool for a job. Describe the task in plain English (e.g. 'check if a token is a rug pull', 'scrape a page to markdown', 'cheap LLM chat') and get a short ranked list of pay-per-call x402 APIs and MCP servers side by side: what each does, exactly how to use it, price, reliability from our own checks, and why it fits. Neutral: ranking is never paid, and you call and pay each tool directly. Use to find pay-per-call APIs and MCP tools from any provider for a task. Answers may be cached for up to 10 minutes. Price: $0.002 USDC per call (x402 or prepaid credits); failed calls are free. In the free trial.
• gas — What a send, token transfer or swap costs right now on each major chain, and which chain is cheapest. Live network fees on Ethereum, Base, Arbitrum, Optimism, Polygon, BNB, Avalanche, Solana and Bitcoin: slow/standard/fast fee rates, congestion and trend, and the USD cost of a send, token transfer and swap on each, plus the cheapest chain now. Pick chains with chains (e.g. base,solana); all by default. Use to pick the cheapest chain, or a good moment, to send or swap. Answers may be cached for up to 15 seconds. Price: $0.002 USDC per call (x402 or prepaid credits); failed calls are free. In the free trial.
• markdown — Convert any web page, PDF, Word (.docx), Excel (.xlsx/.xls), OpenDocument, CSV or XML URL to clean, LLM-ready Markdown, up to 10 MB: returns title, Markdown, links (web pages) and character count. url is the document address (https://...). Private addresses are blocked; failures are not charged. Use to read one web page or document (PDF, Word, Excel) as Markdown. Answers may be cached for up to 10 minutes. Price: $0.002 USDC per call (x402 or prepaid credits); failed calls are free. In the free trial.
• search-answer — Ask anything: it searches, reads the best pages and writes a short answer with citations. ★ One of our best tools. Web search with a cited answer for AI agents (RAG), open-sources tier: finds pages via Wikipedia, Hacker News, DuckDuckGo and Stack Overflow, reads the top ones (robots.txt respected) and writes a concise cited answer, plus the pages read. q is your question; pages (1-5, default 3) is how many pages to read. Moves to full-web search (and price) when a premium index is connected. Use when you want a written answer with sources, not a list of links. Instead, use `search-open` ($0.002) for links only. Answers may be cached for up to 30 minutes. Price: $0.015 USDC per call (x402 or prepaid credits); failed calls are free. Paid only (not in the free trial). Try search-open free first.
• search-open — Fast, cheap search across Wikipedia, Hacker News, DuckDuckGo and Stack Overflow. ★ One of our best tools. Web search API for AI agents, lowest-cost tier: Wikipedia articles, Hacker News stories, DuckDuckGo instant answers and Stack Overflow questions (quota permitting) in one call, merged, de-duplicated and ranked, with titles, links, snippets and dates. Good for facts and for tech news and discussion (Hacker News). q is the search words; limit caps results; detail=brief trims fields. Use for quick facts, tech and programming lookups where results up to an hour old are fine. Instead, use `search-answer` ($0.015) for a written answer with citations. Answers may be cached for up to 1 hour. Price: $0.002 USDC per call (x402 or prepaid credits); failed calls are free. In the free trial.
• token-price — Live price of any token on Base, Solana, Ethereum, BNB, Arbitrum, Polygon, Optimism or Avalanche from DEX data: USD price, total liquidity, 24h volume, buys/sells, price change (5m/1h/6h/24h), market cap, FDV, pair count and pool age. For a full rug-pull check use token-safety. Use when you only need a token's current price, liquidity and volume. Instead, use `token-safety` ($0.02) to check it is safe. Answers may be cached for up to 30 seconds. Price: $0.005 USDC per call (x402 or prepaid credits); failed calls are free. In the free trial.
• token-report — Everything about a token in one answer: safety, who holds it, liquidity, price action, whales and smart money, and news buzz, with an overall rating. ★ One of our best tools. Full token report, live: safety (honeypot, taxes, owner powers), holders (concentration, exchange/fund/locked share), liquidity (depth, locks, pools), price action, social/news buzz and smart-money signals (trade sizes, buy pressure, exchange and fund flows, whale transfers), with an overall rating and headline, in one call. 8 chains incl. Base, Solana, Ethereum. Use for a complete, live picture of one token before a decision. Instead, use `token-safety` ($0.02) for the scam check only. Answers may be cached for up to 20 seconds. Price: $0.10 USDC per call (x402 or prepaid credits); failed calls are free. Paid only (not in the free trial). Try token-safety free first.
• token-safety — Is this token safe to buy? Honeypot and sell-tax test, holder concentration, liquidity and locks, mint/freeze powers, proxy/upgradeable contract, with a 0-100 safety score and plain-English flags. ★ One of our best tools. Rug-pull and scam check for any token on Base, Solana, Ethereum, BNB, Arbitrum, Polygon, Optimism or Avalanche. Checks honeypot, buy/sell tax, mint/freeze authority, hidden owner, proxy, blacklist, holder concentration, locked liquidity, pool age and market data, then gives flags, a 0-100 safety score and a verdict. Use first for a quick verdict on whether one token (by contract address) is a scam or rug pull. Instead, use `token-report` ($0.10) for holders, whales, liquidity and news too. Answers may be cached for up to 10 minutes. Price: $0.02 USDC per call (x402 or prepaid credits); failed calls are free. In the free trial.
• topup — Start a prepaid-credits top-up so every tool keeps working after the free trial (20 free calls a day). Use when a tool says the free calls or credit have run out, or the user asks to buy credit; to see what's left, use `balance` instead. Creates a payment request only: it returns the exact USDC amount and address to send (Base or Solana) and a QR-code page, and nothing is charged unless the user sends that payment, which is credited about 15 minutes after it confirms. Without a key in the server URL it also creates an API key (shown once) and the new server URL to use; with one (e.g. a free key), it tops up that same key, so your connector URL stays the same, and reuses a top-up still waiting for payment. Free to call.
• wallet-risk — Before you pay, trade with or accept funds from an address: sanctions, hacks, phishing, mixers and scam links, with a clear verdict. Screens a wallet or contract address before you pay, trade with or accept funds from it: sanctions (OFAC), theft/hacks, phishing, money laundering, mixers, scams, fake tokens and malicious contracts, with a verdict (high-risk, suspicious, clean). Base, Ethereum, Solana, BNB, Arbitrum, Polygon, Optimism, Avalanche. Use before sending funds to, or accepting funds from, an address. Answers may be cached for up to 10 minutes. Price: $0.005 USDC per call (x402 or prepaid credits); failed calls are free. In the free trial.