Snapshots and extraction - Scrinly API

Use GET or POST /render/snapshot to return structured page data. Snapshots block images, fonts, media, and—unless style extraction is requested—stylesheets to reduce render time.

curl --request POST "https://api.scrinly.com/render/snapshot" \
  --header "Authorization: Bearer $SCRINLY_API_KEY" \
  --header "Content-Type: application/json" \
  --data '{
    "url": "https://example.com",
    "markdown": true,
    "links": true,
    "structure": true
  }'

Extraction fields

Option	Result
`markdown`	Readable page content converted to Markdown
`html`	Full rendered HTML after browser execution
`links`	Deduplicated absolute HTTP(S) links
`images`	Image source information
`videos`	Video source information
`fonts`	Sanitized font families and delivery kinds
`colors`	Frequency-ranked color palette
`structure`	Headings, calls to action, sections, and positions
`design`	Tokens, typography, breakpoints, motion, and library signals

Metadata is always returned. An unrequested optional field is omitted rather than returned as an empty value.

Authenticated pages

Pass request headers and cookies in a POST body when the target page requires a session.

{
  "url": "https://portal.example.com/report",
  "markdown": true,
  "headers": { "X-Workspace": "demo" },
  "cookies": [
    { "name": "session", "value": "value", "domain": "portal.example.com", "secure": true }
  ]
}

Keep this payload server-side. Scrinly does not include request headers or cookies in AI design projections.

Synchronous and asynchronous modes

Snapshots are synchronous by default. Set async=true or provide webhookUrl to queue the render and receive a job ID. blueprint=true is always synchronous and rejects async or webhook combinations.