Integrations · API5 min read

Generate Product Descriptions and Photos With an API

If your product catalog lives in your own system, use Cormak to prepare listing copy and images from catalog data. Follow one ceramic mug from a free connection check to content you can review before publishing.

A perfume bottle photographed with studio lighting

1. Check the connection for free

Fetch the OAuth metadata. This request needs no account, token or credits. The response identifies the resource and provides information for configuring authorization.

OAuth metadata request
curl --fail https://mcp.cormak.app/.well-known/oauth-protected-resource/mcp

2. Connect your account and check the price

Connect to Cormak with OAuth, PKCE and the store:use scope. Send JSON with POST to https://mcp.cormak.app/store and pass your token in the Authorization: Bearer header. Keep the token on your server or in a local Postman environment. The account owner can revoke access and set a generation limit.

Balance and pricing request body
{ "action": "account" }

Check available_ai_credits and prices in the response. Reading them is free. Text generation costs 1 AI credit. For an image, read the current price for your model and quality from prices.images. Generation uses the connected account’s balance.

3. Generate the product description

Supply the product name and verified attributes. Choose the fields you want back. This request can charge a credit immediately. Replace the example UUID with a new one before running it for your product.

Paid description request
{
  "action": "fill_text",
  "request_id": "7b467f8a-8d87-4a92-8c8a-43df4bccd680",
  "locale": "en_US",
  "product": {
    "name": "White ceramic mug",
    "attributes": [
      {
        "name": "Material",
        "values": [
          "Ceramic"
        ]
      }
    ]
  },
  "options": {
    "fields": [
      "title",
      "description",
      "image_alt"
    ],
    "tone": "neutral",
    "length": "short"
  }
}

When status is complete, read the requested fields from result, such as description_html. For Russian copy, set locale to ru_RU. Verify the description against the supplier catalog before publishing.

4. Start the photo generation and retrieve the file

Use a new UUID for the image. Set expected_credits to the current model and quality price from account. The value 2 below is an example. image_start returns HTTP 202 when it accepts the request. This example generates from a prompt. If the exact shape matters, supply the original photo through reference and compare the result with the source.

Paid photo generation request
{
  "action": "image_start",
  "request_id": "3278b354-fdd3-4600-a84b-c7dd7f544ba9",
  "mode": "prompt",
  "prompt": "A white ceramic mug on a plain background",
  "model": "nano2",
  "quality": "standard",
  "aspect_ratio": "1:1",
  "expected_credits": 2,
  "use_free_photo": false
}
Free photo status check
{
  "action": "image_status",
  "request_ids": [
    "3278b354-fdd3-4600-a84b-c7dd7f544ba9"
  ]
}

If the status is running, wait and check again. On complete, download the result from the temporary URL. If a request is interrupted, retry it with the same UUID and parameters. On unknown, check the account for the result before starting a new paid operation.

PRACTICE OVER THEORY

Turn the next step into one sentence.

Try Cormak Free 5 AI credits monthly · no credit card required