Below is a compact, practical guide + ready-to-use prompt templates and examples you can drop into ChatGPT (or another LLM) to generate product, category and CMS content for PrestaShop. Use these templates to get consistent, SEO-friendly, import-ready output (JSON or CSV) and to iterate quickly.
Quick principles (how to think about the prompt)
- Start with a role and goal: tell the model what it is and what outcome you want (e.g., “You are an e-commerce copywriter who writes SEO-optimized PrestaShop product descriptions”).
- Provide precise context/data: SKU, brand, technical specs, features, target audience, keywords, tone, length limits.
- Specify the exact output format (JSON, CSV, table, plain text, HTML) and character limits for meta fields.
- Give examples or a single sample output if you need a specific structure.
- Constrain and validate: ask for URL-friendly slug, list of tags, alt text for images, and a short summary for category pages.
- Split complex tasks into smaller steps or request schema-first output (e.g., produce only JSON then ask for HTML).
- For bulk tasks, ask for one-row-per-item structured output (CSV or JSON array).
PrestaShop-specific constraints to include in prompts
- meta_title: ~50–60 characters max (search-friendly)
- meta_description: ~140–160 characters
- link_rewrite (slug): lower-case, hyphen-separated, no special chars
- description_short: 80–300 characters
- description (long): can be HTML, use paragraphs, bullet lists
- image alt: short, descriptive, include main keyword
- tags: comma-separated
- features: key/value pairs
- categories: human-readable full name and short SEO summary
- Provide language/locale and currency if needed
- Specify whether content must be GDPR-friendly (no personal data), legal disclaimers, or compliance notes
Ready-to-use system prompt (put this in “system” or at the top)
You are an expert e-commerce copywriter and SEO specialist for PrestaShop stores. Produce concise, persuasive, keyword-optimized product titles, short descriptions, full HTML descriptions, SEO meta fields, URL slugs, tags, image alt texts, and structured JSON or CSV ready for PrestaShop import. Respect the character limits in the instructions. If any essential input is missing, request it.
Product prompt template (single product — JSON output)
Use this when you have product fields and want a structured import-ready JSON.
User input (replace bracketed fields with real values):
Product data:
- name: [Product name]
- brand: [Brand name]
- model/SKU: [SKU]
- category: [Category name]
- short_specs: [e.g., 8GB RAM, 256GB SSD, 13" display]
- target_audience: [e.g., students, gamers]
- key_keywords: [comma separated primary keywords]
- tone: [e.g., professional, friendly, luxury]
- language: [en-US or fr-FR]
- images: [image1.jpg, image2.jpg]
- any forbidden words: [optional]
Prompt to LLM:
You are an e-commerce copywriter for PrestaShop. Using the product data above, produce a single JSON object exactly matching this schema:
{
"name": string,
"meta_title": string (<=60 chars),
"meta_description": string (140-160 chars),
"link_rewrite": string (slug),
"description_short": string (80-300 chars),
"description": string (HTML allowed, use
,
- , no external links),
"tags": [string],
"features": [{"name": string, "value": string}],
"image_alts": [{"image": string, "alt": string}],
"price_note": string (optional short note if relevant)
}
Constraints:
- Include all provided keywords naturally in meta_title, meta_description, and description_short where possible.
- Generate a short SEO-friendly slug for link_rewrite.
- meta_* must respect length limits; description_short max 300 chars.
- For description (HTML), include 3–5 short paragraphs and a bullet list of key benefits.
- Use the tone specified.
- Output only the JSON, nothing else.
Example product -> example output (for model reference):
Input: name=Eco Travel Mug; brand=GreenCup; short_specs=350ml, stainless steel, keeps hot 6h; keywords=eco travel mug, insulated mug; tone=friendly
Output: { ... } (model reads this for structure)
Category prompt template
User input:
- category_name: [Category name]
- parent_category: [if any]
- keywords: [primary keyword(s)]
- audience: [who buys these]
- tone: [e.g., helpful, authoritative]
- language
Prompt:
You are an SEO copywriter. Create these outputs for the category:
1) meta_title (<=60 chars)
2) meta_description (140–160 chars)
3) link_rewrite (slug)
4) short_description (1–2 sentences for category listing)
5) full_description (HTML, 2–3 short paragraphs, include 3-5 bullet-point benefits and an internal link placeholder like /category/[slug]).
Return JSON:
{"category_name":"", "meta_title":"", "meta_description":"", "link_rewrite":"", "short_description":"", "description":""}
CMS page prompt template (e.g., About us, Delivery)
User input:
- page_type: [About us / Delivery / Returns / Size guide]
- tone: [friendly/formal]
- length: [short/medium/long]
- legal_points: [any mandatory points]
- language
Prompt:
You are a web copywriter writing PrestaShop CMS pages. Produce:
- Page meta_title (<=60 chars)
- meta_description (<=160 chars)
- link_rewrite (slug)
- visible HTML content (H1 + 3–6 paragraphs + H2 subsections if long)
- bullet list of mandatory legal points included
Return JSON: {"meta_title":"", "meta_description":"", "link_rewrite":"", "content_html":"", "legal_points":[...]}
Examples (concrete)
1) Quick product example
Input data:
name: “Aurora 5000 Blender”
brand: “KitchenPro”
SKU: KP-A5000
short_specs: 1200W motor, 1.8L capacity, 6 stainless blades, 3 speed + pulse
target_audience: home cooks, smoothie lovers
keywords: “power blender, 1200W blender, smoothie blender”
tone: friendly, trustworthy
images: aurora-5000-1.jpg, aurora-5000-2.jpg
Prompt to LLM: (use the Product prompt template above)
Expected outputs (shortened example):
meta_title: “KitchenPro Aurora 5000 — 1200W Power Blender”
meta_description: “Blend smoothies, soups and sauces with the KitchenPro Aurora 5000. 1200W motor, 1.8L pitcher, six stainless blades — reliable kitchen power.” (<=160)
link_rewrite: “kitchenpro-aurora-5000-blender”
description_short: 2 lines selling benefits
description: HTML with 3 paragraphs + bullet list of features and benefits
image_alts: [{"image":"aurora-5000-1.jpg","alt":"KitchenPro Aurora 5000 blender front view"}, ...]
2) Category example
Input:
category_name: “Running Shoes”
keywords: “running shoes, breathable running shoes, marathon shoes”
audience: runners and fitness enthusiasts
tone: authoritative
Outputs:
meta_title, meta_description, slug, category short sentence and full HTML description with bullet points on fit, cushioning, traction, and recommended distances.
Bulk / CSV workflow
- Provide CSV rows with basic fields (name, brand, SKU, short_specs, keywords).
- Prompt the model to output a CSV or JSON array with matching PrestaShop import columns (meta_title, meta_description, link_rewrite, description_short, description, tags, image_alts). Example: “Input: CSV rows; Output: CSV with added columns meta_title, meta_description, link_rewrite, description_short, description, tags, image_alts. Provide only the CSV.”
Tips for building better prompts and refining output
- Provide a sample filled product as a reference so model matches s.
- Force checks: Ask the model to return a final line like “VALIDATION: meta_title_len=XX, meta_description_len=YYY” for quick QA.
- If you need HTML, require safe tags only (
,
, - , , ,
).
- For multilingual stores, generate content for one language per run; include language code and locale.
- For product combinations or attributes, ask for a “combinations” JSON array: [{"attribute":"Color","value":"Red","reference":"SKU-RED","quantity":10, "impact_price":0},...]
- When doing SEO-first prompts: supply target keyword and 2–3 secondary keywords, ask model to include them with natural density and avoid keyword stuffing.
- To keep brand voice consistent across many items, put voice in the system prompt and generate a short brand sguide (3–5 rules) then ask the model to follow it.
Prompt examples you can copy-paste (single product, short)
System: You are an expert e-commerce copywriter and SEO specialist for PrestaShop stores.
User:
Product: {name:"[NAME]", brand:"[BRAND]", SKU:"[SKU]", specs:"[short specs]", keywords:"[kw1,kw2]", tone:"[tone]", language:"[en-US]"}
Task: Produce these fields as JSON: name, meta_title (<=60 chars), meta_description (140-160 chars), link_rewrite (slug), description_short (<=300 chars), description (HTML; 3 paragraphs + bullet list), tags (comma list), image_alts (list). Use the keywords naturally and keep tone. Output ONLY JSON.
Common mistakes to avoid
- Leaving fields unspecified (e.g., no target audience). If missing, instruct the model to ask for missing info.
- Expecting perfect slugs — explicitly require slug rules.
- Overly long meta titles or descriptions — enforce character constraints.
- Not providing keyword priorities — give 1 primary keyword and up to 3 secondaries.
Validation and QA after generation
- Check meta lengths and slug characters automatically in a script, or ask model to output lengths in the response.
- Spot-check a sample of generated descriptions for factual accuracy (don’t let the model invent specs or certifications).
- If using for import, sanitize HTML and escape quotes in CSV columns.
Example final prompt for bulk CSV generation (short form)
System: You are an e-commerce copywriter and CSV generator for PrestaShop.
User:
I will paste CSV rows with these columns: name, brand, SKU, specs, keywords, images
For each row generate additional CSV columns: meta_title, meta_description, link_rewrite, description_short, description (HTML), tags, image_alts.
Constraints: meta_title<=60 chars, meta_description 140-160 chars. link_rewrite only lower-case letters, numbers and hyphens.
Return only CSV with original and new columns, one product per row.
If you want, paste a product row now and I will generate a ready-to-import example.
---
If you’d like, paste 2–3 real product rows (or a single product), and I’ll produce a fully populated JSON/CSV row and an HTML product description sample you can import to PrestaShop.
