Below is a practical, ready-to-use guide with templates and examples you can paste into ChatGPT (or another LLM) to generate high-quality PrestaShop product entries, category pages, and CMS pages. It covers what to include in the prompt, constraints to enforce, SEO length limits, structured output formats (JSON/HTML), and QA tips.
Key principles
- Give the model a clear role and goal (e.g., “You are an e‑commerce copywriter creating product content optimized for PrestaShop”).
- Supply all product data available (title, brand, features, specs, materials, dimensions, SKU, GTIN, images, target audience, tone, localization). If data is missing, tell the model to ask a clarifying question rather than invent facts.
- Require a structured output (JSON or CSV-ready fields) so you can import or map into PrestaShop easily.
- Enforce SEO length constraints for meta fields and alt text.
- Specify language and tone, and whether to use British/American spelling.
- Add “Do not claim features not provided” and “Do not invent compatibility” to reduce hallucinations.
Essential info to provide in every prompt
- Product name (raw input)
- Category (e.g., “men’s running shoes”)
- Brand
- Short list of bullet features/specs (material, dimensions, weight, power, color, sizes)
- Target audience and use cases
- Any keywords for SEO to include (primary and secondary)
- Tone/voice (e.g., “friendly, concise, 3rd person, trustworthy”)
- Language and locale (e.g., “English — US”)
- Required output fields and format (JSON with keys matching your PrestaShop fields)
SEO and formatting constraints (use these in prompts)
- Meta title: 50–60 characters (max ~60)
- Meta description: 150–160 characters (max ~160)
- Short description (product summary shown in listing): 1–3 short sentences or 50–160 words (decide)
- Long description: descriptive HTML paragraph(s), use
for features, 150–400 words typical
- Alt text: 8–15 words (max ~125 characters) for images
- Slug: lowercase, hyphens, no special chars (max ~75 chars)
- Avoid marketing exaggeration and unverifiable claims (unless provided)
Recommended system prompt (paste into ChatGPT “system” role)
You are an e‑commerce copywriter who creates SEO-optimized, conversion-focused content for PrestaShop products, categories, and CMS pages. Output must be in valid JSON (or HTML where specified) and follow field length constraints. Do not invent product specifications. If required input is missing, ask a clarifying question.
Product prompt template (JSON output)
Provide this exact prompt with your product values substituted:
You are an e-commerce copywriter. Create product content optimized for PrestaShop using the data below. Do not invent specifications. Output only valid JSON with these fields exactly:
{
"name": "",
"slug": "",
"short_description": "",
"description_html": "",
"meta_title": "",
"meta_description": "",
"categories": ["", ""],
"brand": "",
"features": ["", "", "..."],
"specs": {"Weight": "xxx", "Dimensions": "L×W×H", "Material": "...", "SKU": "...", "GTIN": "..."},
"image_alt": "",
"tags": ["", ""]
}
Constraints:
- meta_title max 60 characters. meta_description max 160 characters.
- short_description: 20–50 words, highlights main benefit and target user.
- description_html: 120–350 words. Use 2–3 short paragraphs and one bullet list () of key features. Include 3–4 SEO keywords naturally (primary, secondary).
- slug: lowercase, hyphens, no special characters.
- image_alt: 8–12 words, describe the product and color if provided.
- Use American English. Tone: professional and friendly. No promotional superlatives unless provided by the input.
Product example (fill-in)
Input:
name: "AeroRun Lightweight Men’s Running Shoes"
brand: "AeroGear"
category: "Footwear > Men > Running"
features/specs: "mesh upper, EVA midsole, rubber outsole, 250 g (size 42), color: black/red, sizes 40-46, SKU: AG-AR-001, 30-day warranty"
target audience: "recreational runners and daily joggers"
primary_keyword: "men’s running shoes"
secondary_keywords: "lightweight running shoes, breathable running shoes"
(Paste the above into the model and substitute fields.)
Category page prompt template
You can use this for SEO-friendly category descriptions that sit on category landing pages.
You are an e-commerce copywriter. Create a category page description for PrestaShop. Output JSON with:
{
"category_name": "",
"slug": "",
"meta_title": "",
"meta_description": "",
"description_html": ""
}
Constraints:
- meta_title: ≤60 chars, meta_description: ≤160 chars.
- description_html: 80–250 words. Start with a short opening paragraph (what the category is for), then include a short paragraph about buying tips, and end with a call-to-action (CTA). Include 2–3 keywords from this list: , .
- Tone: helpful, authoritative.
Category example
Input:
category_name: "Women’s Yoga Pants"
primary_keyword: "women’s yoga pants"
secondary_keywords: "yoga leggings, high-waist yoga pants"
target_audience: "yoga practitioners and athleisure shoppers"
CMS page prompt template (About, Shipping, Returns, FAQ)
Create an About Us page, FAQs, shipping policy, or returns policy. Output the requested page as HTML (no images), with headings (, ,
,
), and keep language clear and friendly. Include legal or factual accuracy note: "Do not provide legal advice."
Example - About Us
You are writing an About Us page for PrestaShop. Company: , founded: , mission: , unique selling points: , tone: warm and professional. Output only HTML with an , two subsections, three paragraphs total, and one bulleted list of values.
Bulk generation / CSV mapping
- Provide a CSV with columns matching the keys in your JSON (name, brand, SKU, features, specs, primary_keyword, etc.).
- Use a prompt that instructs the model to read each row and output one JSON per row, or output as newline-delimited JSON (NDJSON) so you can parse programmatically.
- Example system instruction: “Convert each CSV row to a JSON product object. Output NDJSON; one JSON object per line; do not wrap in a surrounding array.”
Sample bulk prompt (for few products)
You are a product content generator. For each product row given, output one JSON object per line with these keys: name, slug, short_description, description_html, meta_title, meta_description, categories, brand, features, specs, image_alt, tags. Respect the field constraints described earlier. If data for a field is missing, leave an empty string.
Formatting to import into PrestaShop
- If you want direct HTML for the description field, include HTML tags in description_html.
- Keep short_description (the product excerpt) plain text or simple inline HTML.
- Map JSON keys to PrestaShop CSV import columns or use PrestaShop’s webservice API for programmatic creation.
Quality control and anti-hallucination tips
- Always include exact specs (weights, dimensions, materials). Tell the model “Do not make up specs.”
- After generation, validate: check that specs in description_html match the product’s specs field.
- Use few-shot examples inside the prompt (1–2 examples of ideal outputs) to train the LLM on formatting.
- For images, ask the model to generate descriptive alt text and captions but never invent visual details not provided (e.g., “If you don’t provide image details, create a generic alt: ‘Product image’”).
Short example prompts you can copy-paste
Product short (single product)
System: You are an e-commerce copywriter... (use earlier system prompt)
User:
Product:
- name: "AeroRun Lightweight Men’s Running Shoes"
- brand: "AeroGear"
- category: "Footwear > Men > Running"
- specs: "mesh upper; EVA midsole; rubber outsole; weight: 250 g (size 42); color: black/red; sizes: 40–46; SKU: AG-AR-001"
- primary_keyword: "men’s running shoes"
Tone: friendly, professional. Output JSON with keys: name, slug, short_description (20–40 words), description_html (150–250 words, use for features), meta_title (≤60 chars), meta_description (≤160 chars), image_alt (≤125 chars). Do not invent features or specs.
Category short (single category)
System: (same)
User:
Category name: "Home Office Desks"
Keywords: "home office desk, ergonomic desk"
Tone: professional and helpful. Output JSON with category_name, slug, meta_title (≤60 chars), meta_description (≤160 chars), description_html (100–180 words). Include buying tips and a CTA.
Checklist before sending prompt to the model
- Did I give product specs and measurements? If not, note missing fields so the model asks.
- Did I include target keywords and tone?
- Did I specify output format and length constraints?
- Did I instruct the model not to invent facts?
- Have I provided examples of desired output (optional but helpful)?
Final note
Start with strict structure and short examples and iterate. For large catalogs, run a small batch and spot-check before full import. Use NDJSON/CSV mapping and pre-validated CSV columns for PrestaShop imports.
If you want, tell me one product or a sample CSV row and I’ll generate a ready-to-import JSON/CSV line following these rules.
