AI Product Content for CS-Cart

A: Create or edit one product of the same type so that its description represents the result you want. Then use Train AI on that product and use the learned product-type instructions for future products of the same type.

During training, the system analyzes the example instead of requiring you to write a long prompt manually. It can learn the description structure, writing style, naming conventions, technical-information format and the type of information that should be emphasized.

A: The system starts from the product information available in CS-Cart. The most useful identifiers are normally the product name, manufacturer/brand and manufacturer part number (MPN). A clear MPN is especially valuable because it helps distinguish similar models and configurations.

AI uses these identifiers to research the product and collect relevant information before preparing the content. The generated description is therefore not based only on the original supplier text. The exact amount of information found depends on how clearly the product can be identified and what reliable product information is available.

A: First check the product name and manufacturer part number. Remove supplier-specific internal codes from the product name when they can be confused with the manufacturer's model code, and make sure the real manufacturer/brand is clear.

For products with many closely related versions, include the exact model, MPN or other identifying configuration information before running Prepare with AI again. Always review generated technical data before publishing it.

A: Image search is based on product identity rather than a single fixed search phrase. The system can use the product name, manufacturer and manufacturer code/MPN to look for images that belong to the correct model.

It also learns from successful previous searches. When useful images have previously been found on particular websites or source addresses for a manufacturer or product type, those learned sources can be used as additional places to search in future requests. This allows image search to become more targeted over time instead of starting completely from zero for every product.

A: Check whether the product has an exact manufacturer name and MPN/model code. Generic names such as “black laptop” or supplier-only codes provide much weaker identification than a manufacturer model and exact part number.

If the product has several visually different variants, make the identifying variant information clear before searching. The system can use learned source information to improve where it searches, but accurate product identity remains the most important input.

A: Translation always starts from the Native language configured in the add-on. The content can then be translated into up to three additional languages selected in the settings.

This makes the native-language product content the master version. If the source content needs correction, it is normally better to correct the native version first and then generate the translations again.

A: The add-on only offers languages that are active in your CS-Cart store. Enable the required language in CS-Cart first, then return to the mAIndware settings and select it as one of the translation languages. The Native language cannot also be selected as a translation language, and the same translation language should not be selected more than once.

A: No. The purpose of training is to learn the rules and style behind the example, not to copy its product facts. Future descriptions should follow the learned structure and presentation while using the actual information for the product being processed.

This is why training works best with a representative example from the same product type. A good laptop description is useful for teaching the preferred laptop structure; it should not cause specifications from that laptop to be reused for another model.

A: Usually not. Product-Type Instructions are most useful for requirements that genuinely matter to your store: preferred structure, naming rules, important technical fields, information that must be included or avoided, and category-specific conventions.

Over-defining every sentence and formatting detail can make generated content less flexible. The goal is to give AI clear store-specific guidance while still allowing it to adapt to the actual product and the information available for it.

A: AI Product Content uses your own OpenAI API account. Check that the API key saved in the add-on is valid and that API billing is enabled for the corresponding OpenAI account. A ChatGPT subscription does not include OpenAI API usage.

If the key was replaced or revoked, create a new API key, update it in mAIndware AI Content → Settings → OpenAI and save the settings.

A: Keep the recommended default model unless you have a specific reason to use another supported model. Changing the model can affect output quality, processing time and API cost.