P2Lab AI ships with around thirty prompt template types. Every type has a default template you can edit and any number of forks. This page is a reference list — use the linked feature page for hands-on details.
| Type | What it does |
|---|
generate_seo.product | Meta title, meta description, keywords for a product. See Generate SEO. |
generate_seo.category | Same for categories. |
generate_content.product | Name + description (and optionally meta) for products. Supports model / web_search / url_fetch modes. See Generate content. |
generate_content.category | Same for categories. |
generate_alt_text.product | Accessible alt text for product images. See Generate alt text. |
generate_categories.product | Propose categories for a product from its text. See Suggest categories. |
generate_keywords.product | Propose search keywords for a product from its text. See Suggest search keywords. |
| Type | What it does |
|---|
extract_measures | Weight, dimensions, units, pack-unit labels from a description. See Extract measures. |
extract_properties | Matched + new property group suggestions from a description. See Extract properties. |
| Type | What it does |
|---|
translate.product | Translate selected product fields to one or more languages. See Translate. |
translate.category | Same for categories. |
improve.product | Rewrite a product field with a preset (SEO, concise, detailed, brand voice, grammar only, custom). See Improve. |
improve.category | Same for categories. |
unify.product | Apply a format spec to a product field. See Unify. |
unify.category | Same for categories. |
unify_extract | Pre-step for Unify: infer a format spec from sample products. Not a user-facing job type — invoked internally by ContentUnifier. |
| Type | What it does |
|---|
inline.improve | Improve the current text selection in the WYSIWYG. |
inline.improve_chat | Multi-turn refinement of the improve result — used by the Improve content with chat flow in the inline modal. |
inline.translate | Translate the current selection. |
inline.unify | Apply a format spec to the selection. |
inline.describe_from_url | Fetch a URL and write a description from its content. |
inline.suggest_fetch_domain_config | Propose CSS selectors for URL-based fetches on a new domain. |
inline.plain_text_format | Output-format hint appended to other inline.* system prompts when the target field is short plain text (typically a product name). Editable like any other template. |
| Type | What it does |
|---|
edit_image.pins_only | Apply only pin-based edits, preserve the rest of the image. |
edit_image.studio_background | Replace the background with a studio gradient. |
edit_image.outdoor | Outdoor / nature setting. |
edit_image.beach | Beach-at-sunset setting. |
edit_image.wooden | Rustic wooden surface. |
edit_image.minimalist | Clean white minimalist aesthetic. |
edit_image.promo_banner | Promotional banner styling. |
edit_image.promo_banner_editorial | Promotional banner, editorial look. |
edit_image.promo_banner_grunge | Promotional banner, grunge look. |
edit_image.custom | User-written prompt. |
See Edit image with AI.
| Type | What it does |
|---|
image_finder.search | Drives the web search that returns candidate product-page URLs. See Image Finder. |
image_finder.pick_image | Second pass — picks the single best image from each candidate page. |
media_index.analyze | One vision pass per library image: title, alt text, content tags, dominant colour and a semantic category for foldering. See Media Library Organizer. |
| Type | What it does |
|---|
meta.fork_wizard | One-shot generator for the Create prompt fork wizard — proposes a fork-ready system + user prompt from a base prompt and a description of the change. |
meta.fork_brainstorm | Multi-turn brainstorm assistant used by the fork wizard’s Discuss with AI entry — same inputs as meta.fork_wizard plus a running chat history. Shares the connection with meta.fork_wizard. |
meta.learn_prompt | Powers learning suggestions — analyses samples and proposes prompt edits. |
meta.prompt_from_image | Derives an image-edit prompt from a reference image you supply. Used by Add a template. |
meta.refine_instructions | Multi-turn chat assistant that helps you refine the Additional instructions text of an inline AI task — opened from the Chat button in the inline modal. |
convert_to_html | Universal post-processor that converts model output to HTML for HTML-typed fields. |