Gemini
Extract structured data from the Google Gemini web interface - rich citation metadata with snippets, highlights, favicons, and site names - as clean JSON.
highlights - the exact excerpts Gemini pulled from each source to support its answer. No other scraper returns this level of citation detail.Overview
Use this endpoint to track brand citations in Google’s AI ecosystem, research which pages Google’s AI trusts for any topic, and monitor how Gemini presents your brand or competitors. Thehighlights field shows exactly which sentences from each source Gemini used - useful for content optimisation.
Unique features
- Highlights: The exact excerpts Gemini extracted from each source to support its answer
- Source snippets: Full text excerpt from each cited page
- Favicon + website name: Human-readable site identity for each citation
- Markdown response: Full response as markdown via
result_markdown
Example request
Response schema
Includes common response fields plus:Common questions
What are highlights and how are they different from snippet?
snippet is a general text excerpt from the page. highlights are the specific sentences Gemini extracted and used to ground its answer - they show exactly which parts of the source Gemini considered relevant. Highlights are especially useful for content optimisation: if Gemini highlights your content, that’s strong signal for AI citation potential.
How does Gemini compare to Google AI Mode?
Both are Google products but distinct surfaces. Gemini is the standalone AI assistant; Google AI Mode is the AI answer panel within Google Search. They return different citation sets, different source mixes, and serve different user intents. Run both for complete Google AI coverage.Authorizations
Your ScrapeLLM API key passed in the request header.
Query Parameters
The prompt to send to Gemini. Maximum 4,000 characters.
1 - 4000ISO 3166-1 alpha-2 country code. Routes the request through infrastructure in
that region for localised responses. JP and TW are not supported. Defaults to US.
2Skip the response cache and always fetch a fresh result.
Maximum seconds to wait for Gemini to respond. Must be between 10 and 600.
10 <= x <= 600Response
Scrape completed successfully.
Always "gemini".
"gemini"
"done" on success.
"done"
Unique identifier for this scrape job.
"job_abc123"
The prompt that was submitted.
"What brands do marketers recommend for email automation?"
The country the request was routed through.
"US"
Plain-text response from Gemini.
"Marketers commonly recommend Mailchimp, HubSpot, and Klaviyo..."
Credits consumed by this request. Always 3 for Gemini.
3
Total end-to-end request duration in milliseconds.
6521.8
true if this response was served from cache.
false
Markdown-formatted response from Gemini with inline citation markers.
"Marketers commonly recommend **Mailchimp**, **HubSpot**, and **Klaviyo**..."
Sources cited by Gemini. Gemini provides the richest citation metadata of any AI platform — each source includes a snippet, highlighted passages, and site name.
Gemini model variant used (e.g. gemini-2.0-flash).
"gemini-2.0-flash"
Hosted .md URL for this response. Expires after 24 hours.
"https://scrapellm.com/md/job_abc123.md"