Enhancement Response Body

The ad enhancement response has the same body shape across successful enhancement outcomes. Always check both success and status.

Structure

{
  "ad_request_id": "string",
  "ad_response_id": "string",
  "status": "string",
  "success": "boolean",
  "execution_time_ms": "number",
  "enhanced_prompt": "string | null",
  "product_name": "string | null"
}

`ad_request_id`

Type: string
Description: Unique identifier for the enhancement request.

`ad_response_id`

Type: string
Description: Unique identifier for the enhancement result. This value is required for acknowledgment.

`status`

Type: string
Description: Outcome label returned by the API. Expected values are ok, rejected, and no_fill.

`success`

Type: boolean
Description: true only when an ad was successfully injected.

`execution_time_ms`

Type: number
Description: Server-side processing time in milliseconds.

`enhanced_prompt`

Type: string | null
Description: The prompt to send to your LLM. This is null when no ad was injected.

`product_name`

Type: string | null
Description: Name of the product associated with the injected ad flow. This is null when no ad was injected.

Response values by HTTP code

HTTP code	`status`	`success`	`enhanced_prompt`	`product_name`
`200 OK`	`"ok"`	`true`	populated	populated
`201 Created`	`"rejected"`	`false`	`null`	`null`
`202 Accepted`	`"no_fill"`	`false`	`null`	`null`

See Enhancement Status Codes for the full meaning of each response code.

Handling all cases

JavaScript
Python
cURL

const originalPrompt = "What are some good ways to advertise with AI?";

const response = await fetch("https://api.adstract.ai/api/ad-injection/start/", {
  method: "POST",
  headers: {
    "X-Adstract-API-Key": "your-api-key",
    "Content-Type": "application/json",
  },
  body: JSON.stringify({
    prompt: originalPrompt,
    request_context: {
      session_id: "session_001",
      user_agent: "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/134.0.0.0 Safari/537.36",
      user_ip: "203.0.113.24",
    },
  }),
});

let prompt;
if (response.status === 200) {
  const data = await response.json();
  prompt = data.enhanced_prompt;
} else {
  prompt = originalPrompt;
}

import httpx

original_prompt = "What are some good ways to advertise with AI?"

response = httpx.post(
    "https://api.adstract.ai/api/ad-injection/start/",
    headers={
        "X-Adstract-API-Key": "your-api-key",
        "Content-Type": "application/json",
    },
    json={
        "prompt": original_prompt,
        "request_context": {
            "session_id": "session_001",
            "user_agent": "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/134.0.0.0 Safari/537.36",
            "user_ip": "203.0.113.24",
        },
    },
)

if response.status_code == 200:
    prompt = response.json()["enhanced_prompt"]
else:
    prompt = original_prompt

STATUS=$(curl -s -o /tmp/adstract_body.json -w "%{http_code}" \
  https://api.adstract.ai/api/ad-injection/start/ \
  -H "X-Adstract-API-Key: your-api-key" \
  -H "Content-Type: application/json" \
  -d '{
    "prompt": "What are some good ways to advertise with AI?",
    "request_context": {
      "session_id": "session_001",
      "user_agent": "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/134.0.0.0 Safari/537.36",
      "user_ip": "203.0.113.24"
    }
  }')

if [ "$STATUS" = "200" ]; then
  PROMPT=$(python3 -c "import sys,json; print(json.load(open('/tmp/adstract_body.json'))['enhanced_prompt'])")
else
  PROMPT="What are some good ways to advertise with AI?"
fi

Structure​

ad_request_id​

ad_response_id​

status​

success​

execution_time_ms​

enhanced_prompt​

product_name​

Response values by HTTP code​

Handling all cases​

Next steps​