Export structured answer data for your research.
Google AI Mode Scraper
Capture Google AI Mode answer text and available source references for your queries.
$3 / 1,000 results
Illustrative output format
queryquery | answerTextanswerText | referencesreferences | cachedcached | search_metadatasearch_metadata | warningwarning |
|---|---|---|---|---|---|
| best museums in Paris | Paris has one of the richest museum scenes in the world. Whether you want blockbuster masterpieces, stunning architecture, or quiet hidden gems, here are the absolute best museums | [] | false | {"backend":"android-chromium","surface":"ai_mode","total_time_ms":16491,"final_url_has_udm50":true} | Google served this AI Mode answer without any sources (ungrounded answer: no source cards or citation pills on the page). The answer is delivered and charged as normal; filter on s |
About these output fields
| Field | Type | Description |
|---|---|---|
query | string | Query. Optional when absent at the source. |
answerText | string | Answer text. Optional when absent at the source. |
references | array | References. Optional when absent at the source. |
cached | boolean | Cached. Optional when absent at the source. |
search_metadata | object | Search metadata. Optional when absent at the source. |
warning | string | Warning. Optional when absent at the source. |
What you can do
Compare saved results in your own reports or applications.
Set your sources
Open the scraper and enter your own URLs, search terms and limits.
Review & run
Check the cost estimate and confirm the task. Follow its progress in your account.
Use your data
Explore saved results and export CSV, JSON or XLSX. Downloads of saved results are free.
Example input
This is the configured parser input. Replace the example sources with your own and review all limits before running.
{
"queries": [
"best museums in Paris"
],
"gl": "us",
"hl": "en"
}Explore input fields
| Field | Type | Notes |
|---|---|---|
queries | array | Questions to ask Google AI Mode, one per line. Each returns the AI answer and its cited sources. |
gl | string | Two-letter country code for the Google market, for example us or gb. |
hl | string | Interface language code, for example en or de. |
location | string | Optional city or region to localize answers, for example "Chicago, Illinois". |
Run through the API
Create an API key in your account, send the input to the queued endpoint and use the returned task ID to retrieve results. Sending a live request uses credits. Replace the example Idempotency-Key with a fresh unique key for each task; keep that key when retrying the same task.
curl --request POST 'https://extracto.cloud/api/v1/parsers/google-ai-mode-scraper/queue' \
--header 'X-API-KEY: ext_your_api_key' \
--header 'Idempotency-Key: replace_with_a_new_unique_task_key' \
--header 'Content-Type: application/json' \
--header 'Accept: application/json' \
--data-binary @- <<'JSON'
{
"input": {
"queries": [
"best museums in Paris"
],
"gl": "us",
"hl": "en"
}
}
JSON- Best fit
- Starter
- rows / mo
- 16.3K
- Concurrency
- Up to 2
- $ / 1k rows
- $3.00
Use monthly volume to plan your credit budget. Estimates use the base rate; package credit price, options, task minimums and rounding may change usage. Purchased credits are paid once, with no automatic renewal. Execution capacity is shared. API submissions: up to 10/second.
Frequently asked questions
What counts as one result?
One answer. One result is one captured answer. Google may return an answer without citations; the output preserves that warning. Generated answers are not independently verified facts, and this is not a Gemini model API.
Can I use the API and No-Code form?
Yes. Submit a task in the web interface or through the Extracto API, then retrieve its saved results after completion.
Your next dataset starts here.
Choose your sources, review the price and run Google AI Mode Scraper.