Santos Quick Intelligence Audit
activeAI Website Intelligence for determining whether public websites can be discovered, understood, trusted, and used by agents. Eleven paid capabilities use USDC on Base mainnet (eip155:8453) via x402 v2 with no account or traditional API key. QUICK INTELLIGENCE (GET /api/audit, $0.015, synchronous): lightweight single-page fetch-and-parse audit. AGENT READINESS (GET /api/agent-readiness, $0.075,…
Settled via Coinbase.
- Transactions · 30d
- 17
- Volume · 30d
- $0.298
- Unique buyers · 30d
- 3
- Uptime · 30d
- 97.8%
- Latency p50
- 105ms
- Reported calls · 30d
- 1
Endpoints (11 live)
GET/api/audit— Runs audit checks and returns audit results for submitted data or workflows. (0.015 USDC on Base)GET/v1/fetch— Fetch one public URL through a hardened safe-fetcher and get the raw text body plus response metadata (final URL after redirects, status, selected headers, byte count, timing). SSRF-guarded — private, link-local, and cloud-metadata addresses are blocked including via redirects — with a 15s timeout, 2MB cap, and ports 80/443 only. Text formats only (HTML, JSON, XML, feeds, plain text, JS, SVG); read-only, no crawling, no JavaScript rendering. (0.002 USDC on Base)POST/api/audit/batch— Batch Quick Intelligence Audits: up to 50 public URLs for one flat payment. Each URL gets the full single-page audit (performance, SEO, accessibility markup, security headers, website-intelligence score). Per-URL failures are isolated and reported; payment settles only when at least one audit succeeds. (0.5 USDC on Base)POST/v1/extract/structured— Extract structured JSON from one public web page against a caller-supplied JSON Schema. Fetches and cleans the page (SSRF-guarded, readability-isolated Markdown, truncated to 8000 characters), then calls Claude with forced tool-use to populate exactly the fields the schema asks for — never fabricating values not present on the page. The extracted data is validated against the caller's own schema before it is returned. (0.08 USDC on Base)GET/api/agent-readiness— Run a bounded, passive Agent Readiness audit of public machine-facing interfaces. The auditor never authenticates to or pays the audited target, creates target accounts, submits forms, or invokes advertised business tools. (0.075 USDC on Base)POST/v1/summarize— Summarize one public web page with Claude. Fetches the page (SSRF-guarded, readability-isolated Markdown, truncated to 8000 characters), then returns a tight JSON brief: title, ~200-word summary, up to 10 key facts, up to 15 named entities, and a word count. Pass an optional focus to prioritize information relevant to a specific topic. HTML pages only. (0.033 USDC on Base)GET/v1/links— Map every hyperlink on a public page: fetch one HTML URL (SSRF-guarded, 15s timeout, 2MB cap), resolve relative hrefs against the final URL, strip fragments, dedupe, and return up to 200 links with anchor text, internal/external kind, and topic tags (docs, pricing, api, careers, social, feed) plus per-category counts. Payment settles only on a successful map. (0.003 USDC on Base)GET/v1/feed— Parse any RSS 2.0, Atom, or JSON Feed URL into one normalized JSON object: feed metadata (title, link, description, feed URL) plus up to 50 items with id, title, url, published date, summary (500 chars max), and author. Fetched through an SSRF-guarded safe-fetcher — private, link-local, and cloud-metadata addresses are blocked — with a 15s timeout and 2MB cap. Payment settles only on a successful parse. (0.003 USDC on Base)POST/v1/audits— Create one Deep Page Audit job with browser-rendered checks and a bounded compute reservation. Payment reserves one job: validated, enqueued, executed once with configured browser, request, and byte limits. Settles when the job is accepted, not on report completion. Slow, broken, or blocking target sites are completed audits of that behavior, not refund cases. (0.225 USDC on Base)POST/v1/extract— Extract one public web page as clean Markdown: main content isolated readability-style, plus title, description, canonical URL, outbound links, and word count. Read-only single-page fetch — no crawling, no JavaScript rendering. (0.005 USDC on Base)GET/v1/screenshot— Render one public page in real isolated Chromium and get image or PDF bytes back synchronously: png (default), jpeg, or pdf (A4 print); desktop 1366x900 or mobile 390x844; full_page for whole-page height. JavaScript executes, so SPAs and client-rendered pages are captured as a real visitor sees them. Fresh anonymous visitor every render, SSRF-guarded. Settles only when bytes are returned — failures are free. (0.01 USDC on Base)
First seen · last seen · last active