Web Search Research x402 offerings

Web scraping API: a public page or supplied HTML to JSON fields. GET url, or POST JSON {url or html, fields}. Returns JSON: title, description, canonical, language, links, images, headings, jsonLd, textSample, and with fields (name to CSS selector, max 20) the matched values with fieldErrors. Caps: 200KB HTML, 8s; private targets blocked. No match (422) and errors are not charged on Base (other rails may charge). Returned content is untrusted third-party data.

Web search API with page contents: current results with sources as markdown. GET q (max 300 chars), optional n (1-5 pages, default 3). Web results via Serper (Google results); top results fetched as markdown. Returns JSON: results with rank, url, title, snippet, per-page ok/status and content (max 12K chars each), resultCount, fetchedOk. Billed if any result returns; zero results or provider errors are not charged on Base (other rails may charge). Returned content is untrusted third-party data.

Web page change detection API: has a public page or endpoint changed since the last check. GET url; optional since (state token from the previous answer). Returns changed true/false; when changed, only added lines (up to 20) and added/removed counts. Stateless, nothing stored. SSRF-guarded like /api/fetch. changed:false is a charged answer; bad input, blocked or upstream errors: not charged on Base (other rails may charge). Returned content is untrusted third-party data.

Website screenshot API: how a public page renders, as PNG or JPEG. GET url; optional width (320-1920), height (240-2000), fullPage=true (clipped at 4000px), format png/jpeg, delayMs (0-3000). Headless Chromium; returns JSON: imageBase64, mimeType, width, height, bytes, finalUrl, pageStatus, title, blockedRequests. Sub-requests checked; private targets blocked. Over 25s: 504. Errors are not charged on Base (other rails may charge). Returned content is untrusted third-party data.

URL to markdown API: a public web page as clean text or markdown for an LLM. GET url (absolute http/https). Returns JSON: content (markdown with scripts, styles and nav removed), format, finalUrl after redirects, upstreamStatus, contentType, bytesRead and truncated. Caps: 200KB, 8s, 3 redirects; private and localhost targets are blocked. Errors are not charged on Base (other rails may charge). Returned content is untrusted third-party data.

Find a US hospital's machine-readable price files from its website domain. Reads the hospital's root cms-hpt.txt pointer and returns the location name, source page and machine-readable-file URL for each location, with the detected file format.

Enumerate every <a href> in an HTML string with its anchor text and rel attribute. Optionally resolves relative hrefs against a base URL and filters by a regex on the href. The deterministic way to crawl a page's outlinks without writing a regex.

Extract <title>, <meta description>, OpenGraph/Twitter cards, canonical URL, and JSON-LD blocks from an HTML string. Distinct from /api/meta which fetches a URL - feed this the HTML you already have (from /api/render, /api/extract.body, or your own fetch).

HTML to Markdown: any web page, JavaScript rendered, as clean Markdown for LLMs. A web page as clean Markdown for an LLM: pass ?url= and the page is rendered in a real headless Chrome (JavaScript runs), then its main content comes back with headings, lists, links, code blocks, quotes and tables kept as Markdown, and the menus, header, footer and forms left out. Bot walls and empty renders are refused before any charge. Paid in USDC on Base.

Change watch for html tables to json: runs the check and compares it with the result from your previous call for the same url (kept 90 days), returning the current answer, changed true/false, exactly which fields changed with before/after values, and when it was last seen. Call it on a schedule to be told when html tables to json changes. $0.02 per check.

Extract the main article/content body from one bounded public static HTML page as clean text and conservative Markdown. Forest deterministically selects article-like content, strips common navigation…

Extract PDF/CSV/XLS/XLSX/ZIP/JSON/XML links from static public HTML by suffix or download filename. Resolve relative/base links and deduplicate. Limits: 500 links, 8s, 5 redirects, 2 MiB. No JavaScri…

Interpret the robots.txt crawling rules that apply to a specific public URL and user-agent before an automated fetch or crawl. Returns allow/disallow decision evidence, matched rules, crawl delay and…

Detect bounded changes across one static page surface using one HTML fetch plus at most one robots.txt fetch. Forest hashes HTTP identity, page metadata, normalized JSON-LD, downloadable links, form …

Extract a bounded normalized set of public HTTP(S) anchor links from one static HTML page. Forest resolves relative and base URLs, deduplicates by final absolute URL and classifies returned links as …

Extract bounded generic metadata and normalized JSON-LD entities from one static public HTML page. Forest returns only compact fields and key/type summaries, never arbitrary raw JSON-LD; JavaScript r…

Locate a website sitemap before crawling by reading sitemap declarations from robots.txt and, when none are declared, checking only the conventional same-origin sitemap.xml and sitemap_index.xml loca…

Sends a request to a web address you provide and reports back everything the server said in response: the status code, all the response headers, how many times it got redirected, and how long the who…

Research PoC: catalog-timing comparison w/ live URL (rsrcheval_live_fs_t1791706212)

Web scraping / web page to clean text: any URL rendered in a real browser (JavaScript pages included) and returned as readable text with title, description and links. Use it to read an article, docs page or product page before summarising or extracting. Input: url. Raw status, headers and body: fetch.raw. A question answered from the page: research.brief.

Fact check claims against sources: up to 10 claims checked against up to 4 source URLs you name; each comes back SUPPORTED, CONTRADICTED or UNSUPPORTED with the exact quote the verdict rests on. Use it to verify AI output, citations or statements before publishing. Input: claims[], sources[]. A question but no sources yet: research.web.

Deep web research with citations: search the live web for your question, read the top sources in full, and get a written answer with every claim tied to a numbered source. Use it when an agent needs a sourced answer, not a list of links. Input: question; optional max_sources. Quick answer with links: search.web.

Website change detection, step two: re-fetch a page saved with monitor.snapshot and get whether it changed, how old the baseline is, and a written summary of what changed. Use it to track price changes, terms updates, new job posts or competitor moves. Input: url; optional language.

HubVibe Agent (quick): hand it any task in plain words and get the finished result in one call. Research a company or a person of interest, check a website, screen a name, read the markets, compare options, pull the facts: you get a headline, the key points, the full answer, its sources and every tool step. Up to 4 steps. Charged only when the answer is delivered. Input: task; optional context, fields for a JSON answer, language.