Datacrawl: search, fetch, and extract for AI agents

One credit-metered API. Web and news search, block-aware page fetching with optional JavaScript rendering, and clean structured extraction — the three tools every agent pipeline needs, behind one key.

Start free — 500 credits/mo, no card Read the docs

/v1/search

Commercially licensed web + news search, normalized results with attribution.

/v1/fetch

Escalating fetch cascade: plain HTTP first, a real browser when a page needs it. Honest per-attempt reporting, never a silent failure.

/v1/extract

Readable article text, JSON-LD, or rule-based fields — from a URL we fetch or HTML you already have.

Thirty seconds to first result

curl -s https://datacrawl.dev/v1/search \
  -H "Authorization: Bearer ak_YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{"q": "latest EU AI act obligations", "count": 5}'

Works from anything that speaks HTTPS — or add the datacrawl-mcp package and your agent gets all three as MCP tools.

Pricing

Plan	$/mo	Credits/mo
free	$0	500
hobby	$9	3,000
pro	$49	25,000
scale	$199	120,000

Credit costs: search = 1, plain fetch = 1, rendered fetch = 5, extract = 2+. Every response reports X-Credits-Remaining. Full table and honest capacity notes in the pricing docs.