AI Video Generation API for Beauty and Cosmetics Brands
A cosmetics brand launching a foundation line in 20 shades has product photos from its manufacturer but no application footage — no video of the formula blending into skin, no swatch-swipe clip for the product page. Booking a beauty videographer and a model for a 20-shade application shoot costs far more than most DTC brands budget for a single launch, and the shoot has to happen before the storefront can show the product in use at all. Turning existing product photos into short application-style clips via API answers the "what would this cost" question before committing to a shoot.
Direct answer: GenRelay's video models bill either per second of output or per generation, and a 5-second product demo clip costs between $0.05 and $0.30 depending on the model and resolution — low enough to generate a swatch or application clip for every shade in a launch rather than reserving video for a handful of hero shades.
How Much Does a Single Beauty Demo Clip Cost?
A 5-second clip's cost depends entirely on which model and resolution generate it, since per-second models scale with duration while per-generation models charge a flat fee regardless of length. As of September 2026, GenRelay's video pricing: Grok Imagine 1.0 (text-to-video) $0.010/s, Grok Imagine 1.5 (image-to-video) $0.022/s, Veo 3.1 Lite $0.060/s at 720p or $0.120/s at 1080p, and Gemini Omni Flash a flat $0.10 per generation at 720p or $0.15 at 1080p.
| Model | Mode | Billing | 5s clip cost (720p) | Native audio |
|---|---|---|---|---|
| Grok Imagine 1.0 | Text-to-video | Per second | 5 × $0.010 = $0.050 | No |
| Grok Imagine 1.5 | Image-to-video | Per second | 5 × $0.022 = $0.110 | No |
| Veo 3.1 Lite | Image-to-video | Per second | 5 × $0.060 = $0.300 | Yes |
| Gemini Omni Flash | Image-to-video | Flat per generation | $0.10 (720p) | No |
For a product photo of a foundation bottle or lipstick tube with no existing application footage, image-to-video is the relevant mode — it animates the product photo itself into a swatch swipe or texture close-up rather than generating a scene from a text description alone.
Which Model Fits Beauty Tutorial and Demo Clips?
Grok Imagine 1.5's lower per-second rate fits a full-shade rollout where every shade gets a swatch clip regardless of expected sales volume; Veo 3.1 Lite's native audio is a better fit for a smaller set of hero application clips meant for paid social, where ambient sound (a brush tap, a lid click) adds to the scene rather than playing silently.
Gemini Omni Flash's flat per-generation price becomes cheaper than Grok Imagine 1.5 once a clip runs past roughly 4.5 seconds at 720p, which matters for brands standardizing on a fixed clip length across a shade range rather than varying duration per SKU.
How Do I Generate a Swatch or Application Clip From a Product Photo?
Submit the product photo with a prompt describing the intended motion — a swatch swipe or a formula blending on skin reads better than trying to animate only the packaging in isolation.
import requests
import time
API_KEY = "YOUR_GENRELAY_KEY"
BASE_URL = "https://genrelay.ai/v1"
def submit_job(image_url, prompt, duration=5, model="grok-imagine-1.5"):
r = requests.post(
f"{BASE_URL}/videos/generations",
headers={"Authorization": f"Bearer {API_KEY}"},
json={
"model": model,
"mode": "image-to-video",
"image_url": image_url,
"prompt": prompt,
"duration": duration,
},
timeout=30,
)
return r.json()["id"]
def poll_job(job_id, interval=5, timeout=180):
elapsed = 0
while elapsed < timeout:
r = requests.get(
f"{BASE_URL}/videos/generations/{job_id}",
headers={"Authorization": f"Bearer {API_KEY}"},
)
data = r.json()
if data["status"] in ("completed", "failed"):
return data
time.sleep(interval)
elapsed += interval
raise TimeoutError(f"job {job_id} did not finish in {timeout}s")
job_id = submit_job(
"https://cdn.example.com/products/foundation-shade-12.jpg",
"Foundation liquid swatch swiping across bare skin, blending smoothly into an even finish, "
"soft natural lighting, close-up macro shot",
)
result = poll_job(job_id)
print(result.get("output_url"))
What Does a Full Shade Range Rollout Cost?
A 20-shade launch at 5-second clips via Grok Imagine 1.5 costs 20 × 5 × $0.022 = $2.20. Adding a hero tier of 5 application clips through Veo 3.1 Lite at 720p with audio adds 5 × 5 × $0.060 = $1.50. A combined rollout — full shade range plus hero tier — totals $2.20 + $1.50 = $3.70 for a 20-shade launch with video.
from concurrent.futures import ThreadPoolExecutor
SHADE_RANGE = [
("shade-06", "https://cdn.example.com/products/foundation-shade-06.jpg"),
("shade-12", "https://cdn.example.com/products/foundation-shade-12.jpg"),
("shade-18", "https://cdn.example.com/products/foundation-shade-18.jpg"),
]
PROMPT_TEMPLATE = (
"Foundation liquid swatch swiping across bare skin, blending smoothly into an even finish, "
"soft natural lighting, close-up macro shot"
)
def generate_and_wait(entry):
shade, image_url = entry
job_id = submit_job(image_url, PROMPT_TEMPLATE)
return shade, poll_job(job_id)
with ThreadPoolExecutor(max_workers=3) as pool:
results = list(pool.map(generate_and_wait, SHADE_RANGE))
for shade, r in results:
print(f"{shade}: {r.get('output_url')}")
| Package tier | Composition | Cost |
|---|---|---|
| Full shade range clips (Grok Imagine 1.5) | 20 × 5s | $2.20 |
| Hero application clips (Veo 3.1 Lite, 720p) | 5 × 5s | $1.50 |
| Combined rollout | Both above | $3.70 |
| Flat-rate alternative (Gemini Omni Flash) | 20 × $0.10 | $2.00 |
At $3.70 for a 20-shade launch with a hero tier included, adding a swatch clip to every shade's product page costs less than a single hour of a beauty videographer's time, without scheduling a full application shoot around a launch calendar. See the image-to-video pipeline guide for the async polling and error-handling pattern this batch relies on at larger shade-range sizes.
Internal Links
FAQ
Can the API show a specific model's skin tone in the application clip?
It generates a plausible skin surface in motion from the prompt and reference image provided; matching a specific model or skin tone reliably works best with a clear reference photo as the image input.
Is per-second or per-generation billing cheaper for shade-range clips?
It depends on clip length — Grok Imagine 1.5 stays cheaper than Gemini Omni Flash's flat $0.10 (720p) up to roughly 4.5 seconds; past that, the flat rate wins regardless of duration.
Does Veo 3.1 Lite's audio track require a separate API call?
No. Audio generates synchronized with the video in the same request when using Veo 3.1 Lite — there's no separate audio generation step or additional endpoint.
How fast can a 20-shade batch finish?
Individual clips typically complete within a few minutes; a 20-clip batch submitted with bounded concurrency finishes well within a same-day turnaround for a launch deadline.
Is there a free tier to compare models before committing to a full shade range?
Yes. GenRelay includes free credits on signup, enough to generate sample clips across Grok Imagine 1.5 and Veo 3.1 Lite before choosing a model for the full rollout.
As of September 2026. Pricing subject to change — verify current rates at genrelay.ai.