AI Query

Ask natural language questions about any website. Orsa scrapes the site, passes content to an LLM, and returns structured answers.

Endpoint: POST /v1/brand/ai/query Credits: 10 per request

Parameters

Parameter	Type	Required	Description
`domain`	string	Yes	Domain to analyze
`data_to_extract`	string	Yes	Natural language description of what to extract
`url`	string	No	Specific URL to analyze (default: domain homepage)
`response_format`	string	No	`text` or `json` (default: `json`)

Request Body

{
  "domain": "stripe.com",
  "data_to_extract": "pricing plans with names, prices, and key features",
  "response_format": "json"
}

Response Schema

{
  "success": true,
  "data": {
    "domain": "stripe.com",
    "query": "pricing plans with names, prices, and key features",
    "result": {
      "plans": [
        {
          "name": "Integrated",
          "price": "2.9% + 30¢ per transaction",
          "features": ["Full platform", "135+ currencies", "Instant payouts"]
        }
      ]
    },
    "source_urls": ["https://stripe.com/pricing"],
    "model": "gpt-4o",
    "tokens_used": 1240
  },
  "credits_used": 10
}

Code Examples

cURL

curl -X POST "https://api.orsa.dev/v1/brand/ai/query" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "domain": "stripe.com",
    "data_to_extract": "pricing plans with features",
    "response_format": "json"
  }'

TypeScript

const result = await client.ai.query({
  domain: 'stripe.com',
  dataToExtract: 'pricing plans with names, prices, and features',
  responseFormat: 'json',
});
 
console.log(result.result);      // Structured JSON response
console.log(result.sourceUrls);  // Pages that were analyzed

Python

result = client.ai.query(
    domain="stripe.com",
    data_to_extract="pricing plans with names, prices, and features",
    response_format="json",
)
 
print(result.result)
print(result.source_urls)

Tips

Be specific in data_to_extract — "pricing plans with prices" works better than "tell me about the company."
Use response_format: "json" for programmatic consumption and "text" for human-readable summaries.
The AI may navigate to subpages (like /pricing) if the homepage doesn't contain the requested data.
Results are not cached by default since queries are freeform.

Prefetch by Email Extract Products