curl -X POST https://eachsense-agent.core.eachlabs.run/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "X-API-Key: $EACHLABS_API_KEY" \
  -d '{
    "messages": [
      {
        "role": "user",
        "content": "Generate a product video ad: a sleek matte black wireless speaker slowly rotates on a reflective dark surface, studio lighting creating clean highlights on the curved edges. Camera slowly orbits around the product. Minimal, premium aesthetic, dark background with subtle blue accent light, 4K commercial quality"
      }
    ],
    "stream": false
  }'

Using Python (OpenAI SDK)

from openai import OpenAI

client = OpenAI(
    api_key="YOUR_EACHLABS_API_KEY",
    base_url="https://eachsense-agent.core.eachlabs.run/v1"
)

response = client.chat.completions.create(
    model="eachsense/beta",
    messages=[{
        "role": "user",
        "content": "Generate a product video ad: a sleek matte black wireless speaker slowly rotates on a reflective dark surface, studio lighting creating clean highlights on the curved edges. Camera slowly orbits around the product. Minimal, premium aesthetic, dark background with subtle blue accent light, 4K commercial quality"
    }]
)

print(response.choices[0].message.content)

With Reference Image

Provide your product photo to animate it:

curl -X POST https://eachsense-agent.core.eachlabs.run/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "X-API-Key: $EACHLABS_API_KEY" \
  -d '{
    "messages": [
      {
        "role": "user",
        "content": [
              {"type": "text", "text": "Create a product video ad from this product image. The product smoothly descends into frame from above and lands softly on a marble surface. Camera pushes in slowly for a close-up. Dramatic studio lighting, premium feel, commercial video quality."},
              {"type": "image_url", "image_url": {"url": "https://example.com/my-product.jpg"}}
            ]
          }
    ],
    "stream": false
  }'

Images are sent inside messages using the OpenAI multimodal content format. Maximum 4 images per request.

Streaming

Set "stream": true for real-time SSE responses, or "stream": false for complete result in a single response. Streaming is useful for showing progress in UIs; non-streaming is simpler for scripts and automation.

Video Ad Formats

Format	Aspect Ratio	Duration	Platform
Instagram Reel	9:16 vertical	15-30s	Instagram, TikTok
YouTube Pre-Roll	16:9 horizontal	6-15s	YouTube
Website Hero	16:9 horizontal	5-15s (loop)	Shopify, brand sites
Story Ad	9:16 vertical	5-15s	Instagram, Snapchat
Square Social	1:1 square	15-30s	Facebook, Instagram feed

Shot Types for Product Videos

Shot	Motion	Effect
Product Orbit	camera circles the static product	shows all angles, premium feel
Drop Reveal	product falls/descends into frame	dramatic entrance
Push-In	camera moves toward product	draws attention to detail
Turntable	product rotates on a platform	360-degree view
Lifestyle Cut	product in use by a person	context and aspiration
Macro Detail	extreme close-up gliding over surface	texture and craft
Exploded View	components separate and reassemble	engineering / feature callout

Prompt Engineering Tips

Prompt Structure

"Generate a product video ad:" + [product description] + [motion/animation] + [camera movement] + [setting and lighting] + [mood and quality]

Motion Keywords

slowly rotates, gently descends, smoothly slides into frame,
orbiting camera, push-in close-up, tracking alongside,
parallax movement, subtle floating hover, product assembling

Lighting for Video

studio three-point lighting, dramatic rim light,
soft diffused top light, colored accent backlight,
natural golden hour, high-key bright and clean,
low-key dark and moody, volumetric light beams

Examples

Skincare Product Reveal

curl -X POST https://eachsense-agent.core.eachlabs.run/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "X-API-Key: $EACHLABS_API_KEY" \
  -d '{
    "messages": [
      {
        "role": "user",
        "content": "Generate a product video ad for a luxury skincare serum: a frosted glass bottle with a gold dropper cap descends slowly through soft mist onto a white marble surface. A single water droplet falls onto the marble and ripples outward. Camera pushes in to a close-up of the dropper releasing a golden serum drop. Soft diffused lighting, clean white and gold palette, luxurious spa aesthetic, 9:16 vertical format for Instagram."
      }
    ],
    "stream": false
  }'

Tech Product Launch

curl -X POST https://eachsense-agent.core.eachlabs.run/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "X-API-Key: $EACHLABS_API_KEY" \
  -d '{
    "messages": [
      {
        "role": "user",
        "content": "Generate a product video ad for a smartwatch: starts with a dark screen, a beam of light sweeps across revealing the watch face glowing on a dark surface. Camera orbits around the watch showing the slim profile and metal band. The screen lights up showing a fitness dashboard. Sleek modern tech aesthetic, dark background with blue accent lighting, cinematic lens flare, Apple-style product commercial, 16:9 landscape."
      }
    ],
    "stream": false
  }'

Food Product Social Ad

curl -X POST https://eachsense-agent.core.eachlabs.run/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "X-API-Key: $EACHLABS_API_KEY" \
  -d '{
    "messages": [
      {
        "role": "user",
        "content": "Generate a product video ad for artisanal chocolate bars: a hand slowly unwraps gold foil to reveal a dark chocolate bar, then breaks it in half with a satisfying snap. Close-up of the cross-section showing rich texture. Melted chocolate drizzles in the background. Warm moody lighting, dark wood surface, premium food photography video style, slow motion detail, 1:1 square format."
      }
    ],
    "stream": false
  }'

Sneaker Drop Announcement

curl -X POST https://eachsense-agent.core.eachlabs.run/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "X-API-Key: $EACHLABS_API_KEY" \
  -d '{
    "messages": [
      {
        "role": "user",
        "content": "Generate a product video ad for a sneaker release: a pair of white and neon green sneakers drops from above and lands on a concrete surface with a puff of powder. Camera orbits around the shoes at ground level, showing the sole tread and knit upper texture. Urban graffiti wall in the blurred background. Streetwear aesthetic, bass-heavy energy feel, dynamic camera, 9:16 vertical for TikTok."
      }
    ],
    "stream": false
  }'

Multi-Scene Ad Workflow

PRODUCT="a premium leather wallet in cognac brown"
STYLE="warm studio lighting, dark background, luxury commercial quality"

SCENES=(
  "Scene 1 — Reveal: the wallet slowly slides into frame on a dark leather surface, camera positioned low, dramatic side lighting"
  "Scene 2 — Detail: extreme close-up tracking along the wallet edge showing hand-stitched seams and grain texture"
  "Scene 3 — Feature: the wallet opens to reveal card slots, a person's hand slides a card in and out smoothly"
  "Scene 4 — Lifestyle: a well-dressed man at a restaurant table casually places the wallet beside a glass of whiskey, shallow depth of field"
  "Scene 5 — End card: the wallet centered on a dark surface, camera slowly pulling back, space for text overlay"
)

for SCENE in "${SCENES[@]}"; do
  curl -X POST https://eachsense-agent.core.eachlabs.run/v1/chat/completions \
    -H "Content-Type: application/json" \
    -H "X-API-Key: $EACHLABS_API_KEY" \
    -d "{
      \"messages\": [{\"role\": \"user\", \"content\": \"Generate a product video ad clip: $PRODUCT. $SCENE. $STYLE\"}],
      \"stream\": false
    }"
  echo "=== SCENE ==="
done

Post-Production Assembly

AI generates individual clips. To create a complete ad:

Generate 3-6 scene clips covering reveal, detail, lifestyle, and end card
Sequence in a video editor — add transitions, adjust timing
Add text overlays — product name, features, price, CTA (AI text is unreliable)
Layer music and sound — choose tracks matching the brand mood
Export for platform — correct aspect ratio and duration per platform

Common Pitfalls

Text and logos will not render correctly in video. Add all text in post-production.
Too much motion in one clip — keep each clip focused on one movement or action.
Product inconsistency — describe the product identically across all scenes for visual coherence.
Missing end card space — always generate a final shot with clean negative space for text overlays.
Audio is not generated — plan for music and sound design separately.

ai-product-video-ad

AI Product Video Ad

Quick Start

Using curl