httpkobarupayai

QuickBuild

Fetch a public web page (or take raw text) and return a structured JSON brief made by Llama 3.3 70B: title, language, contentType, summary, keyPoints, named entities with types, topic tags, concrete facts (prices, dates, versions, limits) as label/value pairs, sentiment and audience. Grounded in the page only, no invented facts. Reads up to 24,000 chars of main content (2 MB HTML, 10 s fetch timeout). All params optional: with none it summarizes https://example.com. GET with query params, or POST a JSON body for long text (up to 200,000 chars accepted, first 24,000 analyzed).