Advanced Prompt Editor: Drag & Drop Version

Modified on Thu, Aug 7 at 2:14 PM

This section explains how the Prompt field works - the main tool for creating dynamic instructions for product text generation. It supports attributes and conditions, allows you to build both simple and complex logic, adapts to the selected store, and accounts for localization. This means you can generate accurate, personalized texts for each product with just a few clicks. In this guide, you’ll find everything you need to work with it effectively.  

Prompt creating section

1. Components of the Prompt Configuration Tab

The interface tab includes three main sections:

  • Prompt Field - the text area for editing the prompt with integrated insertion logic.

  • Attributes Section - a list of all available attributes for the selected store.

  • Attributes (if filled) Section - a list of conditions (if-blocks) that can be used for dynamic content management.

You can freely switch between the attributes and conditions sections. Each has its own search input, and search results are saved separately for each section.

Search results

Under the prompt field, there are buttons Save, Save & Preview, and a link to the Batch List. Above the attributes and conditions sections, there is a separate Prompt Template block, which contains:

  • Load - to load a prompt template;

  • Save - to save the current prompt as a template.

When saving via Save & Preview, a unique prompt will be created and shown for each product in the product table, taking into account the available attribute values and applied conditions. Thus, the generated prompts can vary significantly in size and content.


2. Data Density Percent

Each attribute has a data density percent — the percentage of data availability across the entire integration. 



  • Displayed in the tooltip when hovering over the attribute;

  • By default, only attributes and conditions with data density > 0% are shown. Clicking "Show more" reveals all elements; 

  • Used as a guideline: if data density = 0%, the value is missing for most products.

  • If data density = 0 and the value is missing, this line will not appear in the generated text. This is especially important for conditions, as they will not activate if no data is available.


3. Prompt Field

The central part of the interface — this is where you write the logic that forms the final text. The field supports:

  • Insertion of attributes and conditions (via click or drag & drop);

  • Free text editing of any complexity;

  • Formatting: headers, lists, symbols, emojis (™, ®, %, smiling face with smiling eyes  etc.);

  • Tags (<h1>, <h2>, <h3>, <h4>, <h5>, <h6>, <strong>, <b>, <u>, <p>, <a>, <ul>, <ol>, <li>)

  • Combining regular text with dynamic blocks.

  • If the prompt becomes lengthy, a vertical scrollbar will appear for easy navigation.

4. Interaction with Elements (Attributes and Conditions)

Inserting into text:

  • Click to Insert — inserts the element at the cursor position;

  • Drag & Drop — drag and drop it to the desired location;

  • Deletion — an element can be deleted by pressing Backspace or clicking the x on it directly in the field.

Color Coding:

  • Gray Gray Star — inactive (element not inserted into the prompt);

  • Green  Green Star — active attribute;

  • Yellow-orange Orange Star  — active condition;

  • After deletion from the prompt — the color reverts to gray Gray Star.

Reuse:

  • The same attribute or condition can be inserted multiple times in different parts of the prompt.


5. Behavior When Creating the Flow

  • The prompt field is initially empty;

  • The Save, Preview, Batch List buttons are activated only after changes;

  • Saving without modifying the prompt content is not possible;

  • While adding attributes and conditions is not mandatory, it is strongly recommended, especially when generating texts for a set of products:

    • This allows you to personalize content based on the product’s specific characteristics;

    • Avoiding template-like content;

    • Improves the quality and relevance of the texts.

6. Attribute Localization

  • A single integrated store can have several language versions (store locales);

  • Attributes may have different names depending on the store language (e.g., product_name for en-US, product_naam for nl-NL);

  • After selecting the store, localized names will be displayed in the attribute and condition lists;

  • If no name is available for a language version, the default store name (marked with a star * in the list) will be used;

  • You can change the localized name in the integration settings → attribute → locale.




8. Recommendations for Writing a Quality Prompt

Yes, you can simply insert attributes into text — and it will work. But if you want your descriptions to look like the work of an experienced copywriter, here are some tips from us:

✅ Create a clear structure

  • Use headers, paragraphs, lists — everything that helps to visually organize the text.

✅ Always check if data is available

  • If you add an attribute directly, and the value is missing — there will be an empty space in the result. It’s better to wrap everything in an if block.

? Comparison: Bad:

Material: "material" 

Good:

if   $material  Material:  $material  end 

✅  If you use paired tags in prompts, make sure they are closed (</tag>).

✅ Avoid repetition

  • Don’t insert the same value multiple times into different blocks — this overloads the text.

✅ Write "humanly"

  • Imagine you’re not a machine, but a copywriter trying to engage the customer. Add lively details, emphasis, and speak directly to the user.

? Example:

if  $brand    Reliability from brand $brand  — a great choice for your comfort. end 

✅ Check the result

  • Click Save & Preview and see how your prompt works on real products. This will help avoid surprises.

Was this article helpful?

That’s Great!

Thank you for your feedback

Sorry! We couldn't be helpful

Thank you for your feedback

Let us know how can we improve this article!

Select at least one of the reasons
CAPTCHA verification is required.

Feedback sent

We appreciate your effort and will try to fix the article