Explore APIs

Fetch one public web page now and return clean markdown (or plain text) of the main content plus title, description, author, published time, canonical, language and sha256. Charged only on success.

URL health check, $0.001 USDC per call: Is a URL up? Live GET from Cloudflare's edge: final HTTP status, every redirect hop with status and latency, total ms, content-type/length, server, TLS, and if it answers 402 the decoded x402 payment terms (version, network, amount, payTo, bazaar). Unreachable host is still a delivered verdict (up:false). No params = runs the example input.

WhiteMagic hosted recall — agent memory for cross-session continuity: read-only hybrid search (BM25 + vector + associative) over a curated public corpus, providing durable context from prior sessions and the public knowledge store. Keyless discovery (initialize/tools/list) is free; free evaluation keys and a daily no-key trial are available. Paid requests activate session leases. The returned sess

A2A Indirect Prompt-Injection & Jailbreak Firewall (/v1/prompt-guard) — Scans untrusted web pages, emails, MCP responses, or A2A payloads for hidden prompt injections and returns sanitized text. (https://mcp-security.agentdiet.dev/v1/prompt-guard)

Browse recent predictions ordered by most recent first. Filter with ?symbol=btc_usdt, ?status=audited|auditing|pending, ?days=30, ?limit=100 (max 1000, max 20 with ?detail=1). Use ?detail=1 to include full pred_candles arrays (default: stripped — use /decision/:id for single full detail). Response includes total count for pagination.

Decision procedure for individual agent micro-payments: when to auto-pay, deliberate, defer, or escalate a single x402/MPP payment request. Uses amount vs standing-policy cap, merchant trust score, remaining budget, and quote freshness. Returns numbered decision steps, exit criteria, and failure modes.

The Federal Acquisition Regulation, current to eCFR, searchable by meaning. FAR Desk's search tool, one call. A pass (POST https://far-desk.quickbeam.workers.dev/pass) covers every call.

Reach for this tool before picking a specific joke category - lists all available joke types. Returns the full taxonomy of joke categories to browse. Essential for discovering what humor is available before making a selection.

Use when you need the current DNS records of a domain or hostname: resolves the requested record types live over DNS-over-HTTPS (Cloudflare 1.1.1.1 with Google fallback) and returns every answer with TTL, plus the response code (NOERROR/NXDOMAIN) and whether DNSSEC validated. Up to 8 types per call.

A record of autodiscover.<domain>: looks up the autodiscover host under the domain you give (A via Cloudflare DoH) and reports whether it exists, its records and TTL. Quick existence check for common hosts like autodiscover without enumerating subdomains. $0.01 per lookup.

Extract a web page to clean Markdown plus metadata (title, description, links, word count)

Use when you need to discover arXiv e-prints by keyword, title, author or field query: pass any arXiv search_query string (e.g. "transformer attention", "au:vaswani", "ti:\"attention is all you need\"", "cat:cs.LG AND all:diffusion"). Returns up to 25 papers with id, title, summary, authors (names + affiliation without emails), primary category, categories, published/updated, comment, journal ref,

Service: Exa (https://api.paysponge.com/search)

Searches the x402 services indexed by AgentsMap by keyword and returns operators, endpoints and recent paid activity.

Domain & Network Lookup: One-call domain report: registration, DNS, email provider, SPF/DMARC, hosting.

Live Amazon deals feed with current discounts: ASIN, title, deal price, original price, discount percentage, deal badge and end time. Filter by department, brand id, 4-star rating cut-off, min/max price, minimum discount or Prime program; each answer lists the category and brand ids it accepts and where the next page starts. 20 marketplaces. Built for deal aggregation and price-drop alerting.

Extract clean structured content from a public web page

Fetch any public web page or document (HTML, PDF, DOCX, XLSX, CSV) and return clean, LLM-ready Markdown: nav/footer/ads/scripts stripped, absolute links, title/byline/description metadata. Query: url (required), links=1|0, max_chars, engine=auto|readability.

Search copies AfterImage already fetched. Does not hit the live web. Empty hits: GET /v1/page with the URL.

DNS lookup: A/AAAA/MX/TXT/NS/CNAME/SOA/CAA, SPF and DMARC presence