For agents, RAG, and long-context systems

URLs in. Structured context out.

SCRYBE fetches pages, resolves dynamic rendering, extracts semantic structure, falls back to OCR when needed, and returns markdown or JSON that feels deliberate instead of scraped.

Input

web pages, dynamic apps, PDFs, OCR-heavy docs

Output

semantic markdown, structured JSON, job-based workflows

Runtime

Python-first API, browser fallback, Node hidden behind OCR

GET /v1/parse/json?url=https://brahmai.in

{
  "format": "json",
  "document": {
    "metadata": {
      "title": "...",
      "description": "...",
      "content_type": "webpage"
    },
    "markdown": "Title: ...\nMeta Description: ...",
    "chunks": null,
    "summary": null
  }
}