Product description → AI ad
Take a paragraph (or bullet list) describing a product and produce a polished
short-form ad. The ad-generator workflow writes the hook + body + CTA itself
based on the description.
When to use this skill
- The user pastes a product description, spec sheet, or feature list — not a
URL.
- They want hook → benefit → CTA structure (a commercial).
- They have no avatar / talking-head requirement (use
revid-blog-to-avatar-video
for that).
- For a live URL, use
revid-shopify-product-promo.
Inputs
| Field | Required | Notes |
|---|
prompt | yes | The product description (the AI uses it as the brief) |
stylePrompt | no | Optional brand voice notes (e.g. "Apple-like, calm, premium") |
aspectRatio | no | Default 9:16 |
targetDuration | no | Default 22 (s) |
mediaItems | no | If you have product images, pass them in media.provided |
Step-by-step
- Validate
prompt has at least ~30 words (otherwise the ad is too thin).
- Build the payload below; if product images were provided, slot them into
media.provided and set media.useOnlyProvided: false (mix with stock).
- POST
/render.
- Poll
/status.
- Return
videoUrl.
API call template
POST /api/public/v3/render
Host: www.revid.ai
Content-Type: application/json
key: $REVID_API_KEY
{
"workflow": "ad-generator",
"source": {
"prompt": "{PRODUCT_DESCRIPTION}",
"stylePrompt": "{OPTIONAL_BRAND_VOICE_NOTES}",
"durationSeconds": 22
},
"aspectRatio": "9:16",
"voice": { "enabled": true, "stability": 0.55, "speed": 1.05, "language": "en-US" },
"captions": { "enabled": true, "position": "middle", "autoCrop": true },
"music": { "enabled": true, "syncWith": "beats", "trackName": "ad-energetic" },
"media": {
"type": "stock-video",
"density": "high",
"animation": "dynamic",
"quality": "ultra",
"imageModel": "ultra",
"videoModel": "ultra",
"turnImagesIntoVideos": true,
"applyStyleTransfer": false,
"provided": []
},
"options": {
"targetDuration": 22,
"promptTargetDuration": 22,
"summarizationPreference": "summarizeIfLong",
"soundEffects": true,
"addStickers": true,
"hasToGenerateCover": true,
"coverTextType": "hook"
},
"render": { "resolution": "1080p", "frameRate": 30 }
}
ad-generator defaults to higher visual quality than article-to-video because
ads compete on the first second. If credits are tight drop quality to pro.
Examples
Failure modes
| Symptom | Fix |
|---|
| Hook is generic | Make prompt specific. "Wireless earbuds" → meh. "Wireless earbuds with adaptive ANC and 38h battery for $179" → strong. |
| Ad reads like a feature list, not a hook | Add stylePrompt: "Lead with a question or emotional hook. Save specs for the middle. End with the price + a single CTA." |
| Visuals don't match the product | Pass real product images via media.provided: [{ url, type: "image" }]. The AI will weave them in. |
| Voice rushes | Lower voice.speed to 0.95. |
| Too many on-screen stickers | options.addStickers: false. |
See also