Is this the Gemini API?

No. Google's Gemini API runs a model you choose for your own prompts. This endpoint returns what the Gemini app shows a logged-out person, including the web sources it grounded on, which is what you need to see how Gemini presents a topic or brand.

How much does it cost?

3 credits per answer, $3.00 → $1.80 per 1,000 answers depending on the pack. Failed and refused answers are free.

Why is it slower than ChatGPT?

Most Gemini answers take 5 to 30 seconds to finish streaming, longer when Gemini searches the web, and we wait for the full answer and its sources before responding.

Do you log in to the assistant?

No. Every prompt is asked the way a logged-out visitor asks it, in a standard browser, and we never use accounts or anything behind a login. Assistants that ask every logged-out visitor to sign in are not offered.

Why does the answer change between runs?

Assistants don't give the same answer twice: wording, sources and products vary from run to run and by country. For tracking, ask each prompt several times (a batch or a monitor) and count how often you're named or cited, rather than reading one answer.

How do I track my brand in the answer?

Pass track with up to 20 brands and domains. Each comes back in mentions: whether the text names it, whether a cited source is on it, which sources and products match, and how early the text first names it (text_position, 0 to 1).

Can I run hundreds of prompts?

Yes. Send up to 100 prompts per call to POST /v1/batches with endpoint: "gemini" and collect the answers from a signed webhook. AI answers cost the same in the batch lane (no 0.5× discount), because each one holds a browser for 5–30 s.

POST /v1/gemini

Gemini Scraper API

Get the answer Google Gemini gives a logged-out visitor on gemini.google.com, with the web sources it grounded the answer on and whether your brand or domain is named or cited.

Updated Oct 28, 2026

Get a free API key Try it in the dashboard

3 credits per answer (5–30 s), the same in batches 2,500 free credits, no card Failed calls are free

Overview

What the Gemini Scraper API does

POST /v1/gemini opens gemini.google.com as a logged-out visitor in a standard browser and sends your prompt. The response has the answer as text and markdown (citations as [n](link)), the sources Gemini grounded it on with the ones it cited inline marked cited: true, and model, the Gemini model it reports to guests.

Gemini answers take longer than a search: most arrive in 5–30 s, about 12 at the median, and answers that search the web take longer. An answer still being written at 90 seconds is finished in the background: the request fails free with upstream_timeout, and the same request sent again within 10 minutes returns it. country asks as a visitor in that country and language sets the interface language.

track returns mentions for up to 20 brands and domains, so the same prompt set you run on ChatGPT can be compared with Gemini, AI Mode and AI Overviews.

What you get

request object: The normalised request: endpoint, prompt, country, language (not chatgpt), web_search (chatgpt), include_html.
text string: The answer as plain text.
markdown string: The answer as Markdown. Citations are n, where n is the position of a source in sources.
html string: The answer's HTML, only with include_html: true.
model string: The model the assistant reports to logged-out visitors, when it reports one.
sources[] array: position (the n in the citations), title, link, domain, snippet, date, cited (true: cited inline in the answer; false: only listed beside it).
search_queries[] array: The searches the assistant ran to answer (its fan-out), where it shows them.
products[] array: Product cards shown with the answer: position, title, brand, price (as shown), extracted_price, currency (ISO 4217 when unambiguous), rating, reviews, merchant, link, image, description.
places[] array: Local businesses the answer showed, same shape as /v1/maps results.
related_questions[] array: Follow-up questions the assistant suggested.
mentions[] array: Only with track, one per term: term, mentioned (the answer text names it), cited (a cited source matches it), sources (positions of matching sources, cited or listed), products (positions of matching products), text_position (where the text first names it, 0 = start to 1 = end; -1 when it doesn't). See AI visibility tracking.
meta object: request_id, credits_used, cached, cached_at, engine (the assistant), latency_ms, parse_quality.

Request

POST /v1/gemini

cURL

Python

TypeScript

fetch

PHP

curl https://api.serpkite.com/v1/gemini \
  -H "Authorization: Bearer $SERPKITE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"prompt":"How do heat pumps work in cold climates?"}'

Response (illustrative; answers vary from run to run)

200 OK · application/json

{
  "request": {
    "endpoint": "gemini",
    "prompt": "How do heat pumps work in cold climates?",
    "country": "us",
    "language": "en"
  },
  "text": "Cold-climate heat pumps keep heating well below freezing because they use variable-speed compressors and refrigerants that still absorb heat from outdoor air at -15 °C or lower…",
  "markdown": "Cold-climate heat pumps keep heating well below freezing because they use **variable-speed compressors** and refrigerants that still absorb heat from outdoor air at -15 °C or lower [1](https://www.example.gov/heat-pumps-cold-climates)…",
  "model": "3.5 Flash-Lite",
  "sources": [
    {
      "position": 1,
      "title": "Heat pumps in cold climates",
      "link": "https://www.example.gov/heat-pumps-cold-climates",
      "domain": "example.gov",
      "cited": true
    }
  ],
  "search_queries": [],
  "products": [],
  "meta": {
    "request_id": "req_01J8ZK4M6Q2V7",
    "credits_used": 5,
    "cached": false,
    "engine": "gemini",
    "latency_ms": 61200,
    "parse_quality": "ok"
  }
}

Every response also carries X-Credits-Used, X-Credits-Remaining, X-Cache and X-Tokens-Estimate headers. Full endpoint reference

Measured Oct 28, 2026

How Gemini Scraper API behaves on this endpoint

Logged-out Gemini prompts (informational, commercial, local and how-to), each asked as a new visitor.

Answered: Most prompts; a refused answer is an error and costs nothing
Time to full answer: Median about 12 s; most within 5–30 s, answers that search the web at the slow end
Model: Gemini's guest model, named in model

Filters and limits

prompt up to 4,000 characters; one turn per request.
country asks as a visitor in that country; language sets the interface language.
No product cards: products is empty for Gemini.
A refusal or "something went wrong" is reported as an error and not billed.

AI visibility

Is your brand in the answer?

track takes up to 20 brands ("Pipedrive") and domains ("hubspot.com", subdomains match).
Each comes back in mentions: named in the text, cited, the matching sources and products, and how early it's named (text_position, 0 to 1).
A monitor asks the prompt hourly, daily or weekly and reports every answer with its mentions and new_sources, to a signed webhook. Each run costs 3 credits.
Answers vary between runs: ask each prompt several times and track how often you're named. AI visibility guide

mentions (illustrative)

{
  "mentions": [
    {"term": "hubspot.com", "mentioned": true, "cited": true, "sources": [2,5], "products": [], "text_position": 0.08},
    {"term": "Pipedrive", "mentioned": true, "cited": false, "sources": [], "products": [], "text_position": 0.41},
    {"term": "Acme CRM", "mentioned": false, "cited": false, "sources": [], "products": [], "text_position": -1}
  ]
}

POST /v1/monitors

cURL

Python

TypeScript

fetch

PHP

curl https://api.serpkite.com/v1/monitors \
  -H "Authorization: Bearer $SERPKITE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"endpoint":"gemini","prompt":"What are the best CRMs for a 10-person startup?","track":["hubspot.com","Pipedrive","Acme CRM"],"interval":"daily","webhook_url":"https://example.com/hooks/serpkite"}'

Reference

Parameters

The fields /v1/gemini accepts in its JSON body.

Gemini Scraper API parameters
Name	Type	Default	Description
`prompt` required	string	–	The message to the assistant, up to 4,000 characters. Required. (q is not accepted here.)
`country`	string	–	Ask as a visitor in this country (ISO 3166-1 alpha-2): the answer follows the visitor's location.
`language`	string	en	Interface language (e.g. en, de). The answer follows the prompt's language.
`track`	string[]	–	Up to 20 brands ("Acme") or domains ("acme.com", subdomains match), 100 characters each, to check the answer for. Each comes back in mentions. A term with a dot is a domain: it matches sources and product links on that domain; other terms match whole words, case-insensitively.
`include_html`	boolean	false	Also return the answer's HTML in html.
`format`	string	json	markdown returns the answer, its sources, searches, products and a Mentions section as one Markdown document; compact keeps text, sources (n, title, link), search_queries, products and places. Values: json, markdown, compact.
`fields`	string	–	Comma-separated projection, e.g. results.title,results.link,knowledge_graph. Cuts tokens.
`max_age`	integer	–	Accept a cached answer up to this many seconds old, at half price. Answers are cached for at most 1 hour, and only a request with max_age reads the cache.

Full reference, error codes and headers are in the /v1/gemini reference. Queue up to 100 prompts with POST /v1/batches, at the same price per answer.

Use cases

What people build with the Gemini Scraper API

AI visibility (GEO) tracking

Ask the prompts your buyers ask and see whether the answer names you, cites your domain or lists your product.

Share of answer over time

A monitor asks the same prompt hourly, daily or weekly and reports every answer with its mentions and new sources.

Citation research

Find which pages and domains the assistant cites for your category, then fix or pitch the ones that matter.

Agent tool

Give an agent a second opinion from a public assistant through the MCP tool or one SDK call.

Pricing