Raw fetch

Get a web page as HTML, markdown or product data in one call

AI agent instructions

Read the Notte skill and follow its setup instructions when setting up Notte or building browser automation. This is the shared source for CLI setup, authentication, and the browser-to-code workflow.

POST /data/web/fetch loads a URL in a cloud browser and returns the formats you ask for. To extract your own schema with an LLM, use Fetch instead.

Try it in the Console

Try this API in the playground, no code needed.

​

Quick start

curl -X POST https://api.notte.cc/data/web/fetch \
  -H "Authorization: Bearer $NOTTE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"url": "https://example.com", "formats": {"markdown": true}}'

{
  "url": "https://example.com/",
  "markdown": {
    "requested": true,
    "success": true,
    "data": "Example Domain\n\nThis domain is for use in documentation examples..."
  },
  "metadata": { "title": "Example Domain", "language": "en" },
  "cache_metadata": { "status": "miss", "age_ms": 0 },
  "isPartial": false,
  "request_id": "9efce509-100a-4bac-9afe-c38323c84303"
}

Each format you request comes back as { requested, success, data }. Formats you didn’t request have requested: false.

​

Parameters

url · string · required

The page to load.

formats · object · required

Which outputs to return. Set any of these to true:

• html: the page HTML
• markdown: the page as markdown
• product: product data
• parse: parsed Google Maps data. Also set parseParams to {"type": "prebuilt", "parser": "google-maps"}

To get parsed Google Maps data, request parse and set the parser:

{
  "url": "https://www.google.com/maps/place/...",
  "formats": {"parse": true},
  "parseParams": {"type": "prebuilt", "parser": "google-maps"}
}

The result comes back in the parsed field of the response.

markdownParams · object

includeLinks (default true), includeImages (default false), inlineImages (placeholder or preserve).

productParams · object

dedupeImages (default false).

sharedParams · object

How the page is loaded:

• waitFor: how long to wait after load, in milliseconds. Default 500.
• actions: steps to run before extracting, in order:
  • {"do": "wait", "timeMs": 1000}
  • {"do": "scroll", "direction": "down", "amount": "viewport", "maxScrolls": 3}
  • {"do": "perform", "action": "click the Load more button"}
• mainContentOnly: keep only the main content. Default false.
• includeSelectors / excludeSelectors: CSS selectors to keep or drop.
• theme: light or dark. Default light.

maxAgeMs · integer

Maximum age of a cached result. Omit for the default policy, 0 forces a fresh load.

See common parameters for timeoutOpts, zdr and tags, and caching for the cache headers.