Extract

POST /extract

Fetch one or more URLs and return clean, extracted content — HTML (via trafilatura), PDF (via pypdf), and JS-heavy pages (via a headless-Chromium fallback). Only successfully extracted URLs are billed.

Request

curl

curl -X POST https://api.crisps.ai/extract \
  -H "Authorization: Bearer $CRISPS_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "urls": [
      "https://arxiv.org/pdf/2412.19437",
      "https://en.wikipedia.org/wiki/Mixture_of_experts"
    ],
    "extract_depth": "advanced",
    "format": "markdown"
  }'

Python

import os, httpx

r = httpx.post(
    "https://api.crisps.ai/extract",
    headers={"Authorization": f"Bearer {os.environ['CRISPS_API_KEY']}"},
    json={
        "urls": ["https://arxiv.org/pdf/2412.19437"],
        "extract_depth": "advanced",
    },
    timeout=120,
)
print(r.json()["results"][0]["raw_content"][:500])

Node.js

const res = await fetch("https://api.crisps.ai/extract", {
  method: "POST",
  headers: {
    Authorization: `Bearer ${process.env.CRISPS_API_KEY}`,
    "Content-Type": "application/json",
  },
  body: JSON.stringify({
    urls: ["https://arxiv.org/pdf/2412.19437"],
    extract_depth: "advanced",
  }),
});
console.log((await res.json()).results);

Try it

POST /extract Try it

API base URL API key

Request body { "urls": ["https://arxiv.org/pdf/2412.19437"], "extract_depth": "advanced" }

Send request

Body parameters

Parameter	Type	Default	Description
urls required	string | string[]	—	One URL or a list (up to 20).
extract_depth	string	advanced	basic or advanced (deeper fallbacks, higher recall).
format	string	markdown	markdown or text.
query	string	—	If set, chunks are reranked against this query.
chunks_per_source	integer	3	Top-N most relevant chunks per source when query is given.
include_favicon	bool	false	Include each URL’s favicon.
timeout	number	—	Per-URL fetch cap in seconds.

Response

{
  "results": [
    { "url": "https://arxiv.org/pdf/2412.19437", "raw_content": "# DeepSeek-V3 …" }
  ],
  "failed_results": []
}

Field	Description
results[]	Successfully extracted URLs with raw_content (and favicon when requested).
failed_results[]	URLs that could not be extracted — not billed.

Credits: 1 per 5 successful URLs (2 per 5 on advanced), reported in the X-Credits-* response headers.