Grok Scraper

Apify Actor that sends a prompt to Grok as a guest user and returns the parsed chat response to the run Dataset. No Grok account is required.

What it does

1. Opens Grok in a stealth Playwright browser through an Apify proxy in your chosen country.
2. Aligns browser locale, timezone, and language headers with that country.
3. Waits for Grok’s session handshake, types your prompt, and submits it.
4. Captures the POST /rest/app-chat/conversations/new response.
5. Parses the NDJSON stream into structured JSON and saves one record to the Dataset.

On failure (rate limits, timeouts), the Actor retries with a new proxy session and browser profile (up to 10 attempts).

Input

Field	Required	Default	Description
prompt	yes	—	Question or instruction for Grok
country	yes	US	Proxy country and browser locale (see list below)
useResidentialProxy	no	true	Use residential proxies; set false for datacenter
includeHtml	no	false	Add the full Grok page HTML to the output

Supported countries: US, GB, DE, FR, CA, AU, IT, ES, NL, BR, IN, JP, MX, PL, SE

Example:

{
  "prompt": "What is the cheapest electric car?",
  "country": "US",
  "useResidentialProxy": true,
  "includeHtml": false
}

Output

One Dataset record per run:

{
  "text": "Assistant reply with citation tags stripped…",
  "searchQueries": ["cheapest electric car 2026"],
  "sources": [
    {
      "position": 1,
      "url": "https://example.com/article",
      "label": "Article title",
      "siteName": "example.com"
    }
  ],
  "model": "grok-3"
}

If includeHtml is true, an html field is added with the full page source.

Run on Apify

1. Push the Actor: apify push
2. Open the Actor in Apify Console
3. Fill in the input form and start a run
4. Open the Dataset tab for the result

Or from the CLI:

apify login
apify push
apify call -i '{"prompt":"What is the cheapest electric car?","country":"US"}'

Requirements: Apify proxy enabled on your account (residential recommended).

Run locally

npm install
npx playwright install chromium
cp .env.example .env   # add your APIFY_TOKEN

Option A — environment variable:

$APIFY_INPUT='{"prompt":"What is the cheapest electric car?","country":"US"}' npm run start:dev

Option B — input file:

mkdir -p storage/key_value_stores/default
cp .actor/INPUT.example.json storage/key_value_stores/default/INPUT.json
npm run start:dev

Results are saved to storage/datasets/default/*.json (not printed to the terminal).

$cat storage/datasets/default/*.json | jq .

Optional local settings (in .env):

• HEADLESS=false — show the browser window while debugging
• DEBUG=1 — verbose scraper logs

Project scripts

Command	Description
npm run start:dev	Run the Actor locally (TypeScript)
npm run build	Compile to dist/
npm start	Run compiled Actor (dist/main.js)
npm run typecheck	TypeScript check

Tags

Grok scraper, LLM ui chat scraper, stealth grok scraper, anti-bot detection, Gemini scraper, Chatgpt scraper, GEO, SEO, AEO