Home / Tools / Gemini Scraper: Answers & Sources: $5.00/1K
Gemini Scraper for answers, source links, and brand checks
kazkn/gemini-ai-answers-scraper
Gemini Scraper collects answers from the anonymous Gemini web interface and turns each processed prompt into a dataset row with Markdown and plain text, ordered web sources, language, exit-IP country, an optional brand match, and a status. It is for SEO and GEO teams, brand researchers, content analysts, and developers who want prompt results they can export, compare, or send into another workflow.
What you get
| Field | What it contains |
|---|---|
prompt | The trimmed prompt submitted to Gemini. |
answer_markdown | The cleaned answer with Markdown preserved; blank when status is not ok. |
answer_text | The answer without Markdown formatting, ready for text analysis; blank for non-ok rows. |
sources | Ordered source objects with url, title, domain, and position; empty when no usable sources are returned or the prompt fails. |
sources_count | The number of entries in the sources array. |
target_brand_mentioned | A case-insensitive whole-name match: true, false, or null when targetBrand is empty. |
country | The ISO two-letter country code of the exit IP, or null if detection fails. |
language | The language code returned by Gemini, or an empty string if unavailable. |
status | The outcome: ok, empty, blocked, or error. Only ok triggers the result charge. |
scraped_at | The UTC ISO 8601 time when the row was created. |
Download as JSON, CSV or Excel, or read the rows through the Apify API.
What people use it for
Track a brand in buyer questions
Set targetBrand to a company or product and run a fixed prompt list. Filter on target_brand_mentioned, read the matching answers, and inspect their sources. Schedule the same input again when you want a new timestamped comparison.
Study Gemini citations for GEO
Submit the questions your audience asks and export the source URLs, domains, and positions. Group rows by prompt or domain to see which sites appeared in that run. Treat this as observed citation data, not a ranking or visibility score.
Research content and competitors
Collect answers for category, comparison, and problem-solving prompts. Use answer_markdown to retain structure or answer_text for analysis, then review the linked sources before using any claims. Run separate targetBrand inputs when you need literal checks for more than one name.
Feed a data pipeline
Start the Actor with the Apify API and retrieve the run's dataset in Python or another API client. Keep ok rows for analysis and use the remaining statuses to identify prompts that need attention. Store scraped_at with each row so later runs can be compared.
Who it is for
- SEO and GEO teams compare answer wording and cited domains across recurring prompts.
- Brand marketers check for a named brand and review the sources returned with each answer.
- Content researchers gather Markdown and plain-text answers for topic and competitor research.
- Developers and analysts pull timestamped rows through the Apify API for their own pipelines.
How to run it
- Open https://apify.com/kazkn/gemini-ai-answers-scraper on Apify and start a new run.
- Paste your prompts, set maxItems if needed, and optionally enter targetBrand or a proxy configuration.
- Run the Actor. It processes unique, non-empty prompts as separate single-turn requests and writes a status for each processed row.
- Open the run's dataset to review the answers, sources, brand matches, language, country, and timestamps.
- Download the dataset as CSV, JSON, or Excel, or fetch it through the Apify API with the run's defaultDatasetId.
Run it from your code
from apify_client import ApifyClient
client = ApifyClient("YOUR_APIFY_TOKEN")
run = client.actor("kazkn/gemini-ai-answers-scraper").call(run_input={"prompts": ["best crm for startups", "best project management tool for agencies", "best running shoes for flat feet", "top vpn for streaming", "cheapest web hosting for wordpress"], "maxItems": 5, "targetBrand": "HubSpot"})
for row in client.dataset(run["defaultDatasetId"]).iterate_items():
print(row)import { ApifyClient } from "apify-client";
const client = new ApifyClient({ token: "YOUR_APIFY_TOKEN" });
const run = await client.actor("kazkn/gemini-ai-answers-scraper").call({"prompts": ["best crm for startups", "best project management tool for agencies", "best running shoes for flat feet", "top vpn for streaming", "cheapest web hosting for wordpress"], "maxItems": 5, "targetBrand": "HubSpot"});
const { items } = await client.dataset(run.defaultDatasetId).listItems();
console.log(items);curl -X POST "https://api.apify.com/v2/acts/kazkn~gemini-ai-answers-scraper/run-sync-get-dataset-items?token=$APIFY_TOKEN" \
-H "Content-Type: application/json" \
-d '{"prompts": ["best crm for startups", "best project management tool for agencies", "best running shoes for flat feet", "top vpn for streaming", "cheapest web hosting for wordpress"], "maxItems": 5, "targetBrand": "HubSpot"}'Your token is in Apify Console, under Settings, API & Integrations. The input above is the example from the tool's input form.
Questions
Is it legal to scrape Gemini answers?
Legality depends on your jurisdiction and what you do with the collected content. Automated access may conflict with Google's terms, and the Actor is not affiliated with Google. It does not bypass CAPTCHAs or anti-bot mechanisms. Review the terms and laws that apply to your use before running it.
How much does it cost to scrape Gemini answers?
Each row with status ok represents a complete answer and triggers a $0.005 result event, which is $5.00 per 1,000 complete answers. The empty, blocked, and error statuses do not trigger that event. Apify platform usage and optional proxy charges are separate, and Apify's standard Actor-start event may also appear in billing.
How do I export Gemini answers and sources?
Open the run's dataset in Apify to download CSV, JSON, or Excel. For automation, start the Actor and read the dataset through the Apify API or the official Python client. The run response provides the defaultDatasetId used to locate the results.
How many prompts can I process in one run?
You can submit up to 10,000 prompt strings, and maxItems can be set from 1 to 10,000; its default is 1,000. Empty prompts and exact duplicates after trimming are removed before processing. Each prompt is single-turn and can receive up to three attempts. A maximum run budget can stop processing before the full list is finished.
Can I use Gemini Scraper without a Google login?
Yes. It uses the anonymous Gemini web interface and does not ask for Google credentials, account cookies, or a Gemini API key. It is not the official Gemini API, so there is no model selector or conversation history. Google can change the internal web endpoint or require authentication, which can affect availability.
Are sources guaranteed for every Gemini answer?
No. Gemini can return a complete answer without usable web sources, leaving sources empty and sources_count at 0. The Actor preserves and deduplicates the sources it receives rather than creating citations. AI answers and linked pages should be checked independently before you rely on them.
Can I target a specific country or language?
There is no country or language selector. The language field records the code reported by Gemini, and country records the exit-IP location, which can be null if detection fails. You can configure an Apify or custom HTTP proxy, but neither the country field nor a proxy proves how Gemini localized the answer; test your own prompts.
Other tools
Vinted Smart Scraper
kazkn/vinted-smart-scraper
Vinted Smart Scraper pulls public Vinted listings, seller profiles and price data from 26 country sites, from vinted.fr and vinted.de to vinted.co.uk and vinted.com, into structured rows.
LoopNet + Crexi Scraper
kazkn/commercial-real-estate-brokerage-intel
Pulls public commercial real estate listings from LoopNet and Crexi, merges them, removes listings that appear on both, and normalizes price, size and cap rate.
Vinted Turbo Scraper
kazkn/vinted-turbo-scraper
Vinted Turbo Scraper reads the Vinted pages you point it at (searches, brand and category pages, seller wardrobes or single items) and returns each listing as a structured row with price, brand, size, condition, photos and seller.
Vestiaire Collective Scraper
kazkn/vestiaire-collective-smart-scraper
This Apify actor collects public Vestiaire Collective listings and sold-search results, with price, currency, condition, brand, model, images and seller country for each item.