From independent questions to reviewable answers

Gemini Scraper API

Use Gemini Scraper API to collect answers, readable formats, and available citations. Research product recommendations or customer questions, then retrieve each answer through a task you can track.

Submit a question and receive a task ID immediately.

Receive a task ID immediately, then automatically check progress every 5 seconds.

3 credits per question, deducted at submission; failed inputs are automatically refunded. Queries are free.

Retry with unchanged inputs reuses the same task. The API supports batches of up to 20 questions. Maximum processing time: 24 hours; undelivered inputs are refunded. Results retained for 30 days. Oversized results that cannot be stored or delivered are refunded.

GETExample response
https://api.goanyapi.com/api/v1/gemini-scraper/tasks/{task_id}
HTML

Loading saved example…

Run a request to see the response here

Fill in the demo parameters to preview an example response for free. Sign in to send live requests with your own parameters.

How Gemini Scraper API works

  1. 01

    Submit a question

    POST /api/v1/gemini-scraper/tasks

    Send your prompt with an optional supported country and tracking index. The API reserves credits and returns a task ID; that confirms acceptance, not a finished answer.

  2. 02

    Wait for the task result

    GET /api/v1/gemini-scraper/tasks/{task_id}

    Query with the saved task ID at least five seconds apart. Stop when the task reaches a terminal state, and inspect every question’s status before analyzing Gemini answers.

  3. 03

    Keep the answer and evidence

    Store original Gemini responses, prompts, and countries with your analysis for review.

PROMPT → ANSWER → EVIDENCE → YOUR BRIEF

Read a Gemini answer before drawing conclusions

A travel content team requests three cities for a Japan guide. Gemini Scraper API supplies the material; the team decides what needs checking.

prompt
Recommend three cities for a first trip to Japan, with a short reason for each.

Saved example · country: US · index: 1

01Gemini answer, summarized

Tokyo: modern city life. Kyoto: traditional culture. Osaka: food and a lively atmosphere.

02Citations in this saved example

  • citations: []
  • links_attached: []
01Capture the recommendation and reason
data.results[].result.answer_text

Read the Gemini answer text and preserve context around each city or product name. A mention count alone cannot explain why an option was recommended.

02Inspect sources separately
data.results[].result.citations / data.results[].result.links_attached

Read citations when present, and inspect links_attached separately. Empty arrays in this saved example mean no links were returned in those fields; they do not turn a successful task into a failure.

Your analysis → a travel-content brief

Create a city-and-reason table, retain the original answer, and add a fact-checking column. Empty citations do not establish source verification.

Task-query responses place these fields under data.results[].result. The default Tester example displays the saved result object directly, so its JSON starts at fields such as answer_text.

Choose the answer format for your workflow

Gemini responses can include several representations. Use the one that fits your next step; fields can be empty or absent.

  1. 01

    Text for extraction

    data.results[].result.answer_text

    Use answer text to identify cities, brands, or recommendation reasons and retain matching passages for review.

  2. 02

    Markdown for readable briefs

    data.results[].result.answer_text_markdown

    When returned, Markdown helps keep headings and lists in an internal report or an Agent’s research context.

  3. 03

    HTML for answer inspection

    data.results[].result.answer_section_html

    The saved example pairs a reconstructed Gemini interface with JSON. Live results still preview the actual answer-section HTML when present; this is answer content, not a full-page screenshot.

What can you build with Gemini Scraper API?

Connect Gemini responses to a defined research deliverable.

Brand and product research

Build a worksheet of products mentioned in Gemini answers, surrounding reasons, and any citations. Your analysis determines the comparison; the API does not assign a brand visibility score.

Content planning

Use Gemini responses to explore how customer questions are explained. For the Japan example, compare the city suggestions with your guide’s coverage and create a list of claims to verify.

Independent question studies

Collect separate answers for distinct audiences or countries. Preserve each prompt and index, then group the results in your own reporting system without assuming shared conversation context.

Give an Agent a focused Gemini research task

After connecting GoAnyAPI, confirm the submit and query tools are available. The Agent organizes returned data; it does not add a new API capability.

Selected task

Prepare the Japan-guide brief

A small Gemini answer sample with an explicit evidence check.

goanyapi_gemini_scraper_submitgoanyapi_gemini_scraper_task
Example research task
Confirm Gemini submit and query tools are available. Show the cost for one US question: “Recommend three cities for a first trip to Japan, with a short reason for each.” After authorization, submit once; poll at least five seconds apart until a terminal status. Check item status. Return cities, reasons, and citations; mark unsupported claims. Do not automatically create another paid task.

Gemini Scraper API FAQ

They serve different workflows. This endpoint collects Gemini website answers for research. It does not expose model selection or conversational follow-ups; each submitted question is independent.

No. A Gemini answer can arrive without citations or attached links. Preserve missing and empty fields, and verify important claims independently.

Submission reserves the configured rate per question. Successful inputs are charged, failed inputs are refunded, and result queries are free. Current rates appear in the Tester and pricing page.

Retry with the same idempotency key and identical inputs. A new key creates a separate task; changing inputs while reusing a key causes a conflict.

Yes. Check each results[].status, not only the overall task status. Analyze successful items separately and retain failure information in your own records.

Completed results are retained for 30 days. Store them externally for longer research histories. Scheduling, trend analysis, and alerts require your own workflow.