Web Search Research x402 offerings

Copy link Define the structure of your output directly in the API request

Extract all links (hrefs) from any web page as structured JSON — every internal, external and anchor link, deduplicated, returned with a total count. Renders the page in a real headless browser, so it captures links injected after JavaScript runs that fetch-only scrapers miss. For crawling, sitemap building, link audits and site-structure mapping. Keyless, pay-per-call.

Web search for AI agents, powered by Exa neural search: pass a query and receive relevant pages with title, URL, published date, author, relevance score and a text snippet as structured JSON. Pair with /web-extract for full-page content. For RAG retrieval, research, fact-finding and monitoring. Keyless, pay-per-call, cached by query. No matches returns an empty result set.

Turn any URL into clean structured JSON for AI agents and RAG. An LLM reads the fully rendered page (JavaScript executed, SPAs included) and returns its title, a 1-2 sentence summary, key entities (people, orgs, products), and any prices, dates or notable numbers — even on pages with no schema.org or JSON-LD markup, no CSS selectors or schema to write. Just pass a URL. Keyless, pay-per-call.

Quick web answer with citations (Exa upstream): ask one factual or current question and get a concise synthesized answer plus source URLs and publication dates. Optimized for fast lookups and agent loops; use /llm-research for a deeper multi-section research report, reasoning, image analysis, or webpage summarization. Keyless, pay-per-call, cached by query.

Importance-ranked crypto headlines with per-article source links and publish times, plus a sentiment-scored market brief. Regenerated hourly; meta carries the snapshot's age and a fresh/stale verdict.

Assess E-E-A-T signals — expertise, authoritativeness, trustworthiness. Looks for author bio markup, About/Contact pages, editorial policy, external citations, review signals, and other trust indicators that Google uses for E-E-A-T evaluation.

Inspect robots.txt, XML sitemap and crawl configuration. Validates robots.txt syntax, sitemap presence and freshness, correct disallow rules, crawl-delay settings, and sitemap–robots consistency. Returns issues that may block search engine crawlers.

Validate HTML structure, semantic elements and document outline. Checks DOCTYPE declaration, charset, viewport meta, deprecated tags, nesting errors, missing semantic landmarks (main, nav, header, footer), and duplicate IDs.

Score content quality — readability, thin pages and duplicate text. Analyzes reading level (Flesch-Kincaid), detects thin content (under 300 words), near-duplicate pages, keyword stuffing, and content-to-HTML ratio. Returns per-page content quality scores.

Map redirect chains, loops and unnecessary 3xx hops. Traces all redirects site-wide, detects chains longer than 2 hops, redirect loops, mixed HTTP/HTTPS redirects, and www/non-www inconsistencies. Returns a redirect map with hop counts.

Test mobile usability — viewport, touch targets and font sizes. Checks viewport meta tag, tap target spacing (minimum 48×48px), font size legibility (minimum 16px body), horizontal scroll issues, and mobile-specific render problems.

Fetch any article or web page and extract clean readable text stripped of navigation, ads, and boilerplate — returns the main content body, word count, estimated reading time, detected language, and key sentences so agents can read the web without a browser or third-party scraping service. For JS-rendered or bot-walled pages a plain fetch can't read, use /exa/contents.

Fetch a reviewed public page and return its links with resolved absolute URLs and anchor text, the canonical URL, the declared meta robots directive, and how many links leave the host. Relative hrefs are resolved against the page, so a caller never has to guess a base.

All links on a page classified as internal or external, with anchor text, nofollow/sponsored, new-tab flags and the most linked external domains. For SEO, crawling and audits.

Fetch a reviewed public page and return its title, every meta and Open Graph tag, and the canonical URL. No JavaScript is executed and no script content is read. Returns the licence and creator for the content alongside it, so reuse terms travel with the result.

Fetch a reviewed public page and return its HTML tables as arrays of rows and cells. Bounded to ten tables, fifty rows each and twenty cells per row, and every bound that is reached is named in the warnings so a truncated table is never mistaken for a short one.

Split a web page's readable text into ~1024-token chunks with 0-token overlap, breaking on sentence boundaries, for embeddings and RAG. Returns the chunks in order with their approximate token counts, the page title and total chunk count. One preset, no parameters to get wrong. $0.01 per page.

Split a web page's readable text into ~1024-token chunks with 100-token overlap, breaking on sentence boundaries, for embeddings and RAG. Returns the chunks in order with their approximate token counts, the page title and total chunk count. One preset, no parameters to get wrong. $0.01 per page.

Split a web page's readable text into ~2048-token chunks with 0-token overlap, breaking on sentence boundaries, for embeddings and RAG. Returns the chunks in order with their approximate token counts, the page title and total chunk count. One preset, no parameters to get wrong. $0.01 per page.

Split a web page's readable text into ~2048-token chunks with 100-token overlap, breaking on sentence boundaries, for embeddings and RAG. Returns the chunks in order with their approximate token counts, the page title and total chunk count. One preset, no parameters to get wrong. $0.01 per page.

Split a web page's readable text into ~2048-token chunks with 50-token overlap, breaking on sentence boundaries, for embeddings and RAG. Returns the chunks in order with their approximate token counts, the page title and total chunk count. One preset, no parameters to get wrong. $0.01 per page.

Split a web page's readable text into ~256-token chunks with 0-token overlap, breaking on sentence boundaries, for embeddings and RAG. Returns the chunks in order with their approximate token counts, the page title and total chunk count. One preset, no parameters to get wrong. $0.01 per page.

Split a web page's readable text into ~256-token chunks with 100-token overlap, breaking on sentence boundaries, for embeddings and RAG. Returns the chunks in order with their approximate token counts, the page title and total chunk count. One preset, no parameters to get wrong. $0.01 per page.