1
Click Start Free
Choose your product photo and type the name, language and call to action. They are saved into one small launch file on your device.
Free · runs on Google Colab
Turn a product photo into a 9:16 ad video
Add one photo and press Run all. Get a 15-second vertical ad with AI motion, voice-over, captions, background music and a call-to-action ending, made on a free Google Colab T4 GPU.
- No sign-up on this site
- Free T4 GPU
- Your photo never leaves your devices
How it works
Five steps, one notebook, no setup.
2
Open Colab
The button opens the notebook in Google Colab, on the free T4 GPU.
3
Run All
Choose Runtime → Run all. It writes the script, records the voice-over, makes the AI clip and adds captions, music and the ending.
4
Upload Photo
When asked, choose the launch file. It already contains your photo. That is the only click.
5
Download Video
A 15-second 480×832 MP4 downloads automatically, ready for Reels, Shorts and Stories.
First run: 15–25 minutes (the models download once)Later runs: 5–8 minutes
Built for free Colab
Everything is tuned to fit the GPU Google gives away.
Free Google Colab
Runs in your own Colab session. Nothing to install and no server to pay for.
T4 GPU
Sized for Colab's free 16 GB T4, with automatic fallback to a shorter AI clip if memory runs out.
9:16 Ads
15-second 480×832 vertical output, made for Reels, Shorts and other full-screen placements.
Voice, Captions & Music
AI script, Hindi, English or Bangla voice-over, animated captions, background music and an “Order Now” ending, mixed for you and downloaded automatically.
Frequently asked questions
Is it really free?
Yes. The notebook uses Colab's free T4 GPU. Google decides GPU availability and session limits for free accounts, so you may occasionally have to wait or try again later.
Do I need an account?
You need a Google account to use Colab. This website has no accounts, no servers that receive your data and no tracking of its own.
Where does my photo go?
Your browser resizes it and saves it, together with your settings, into a launch file on your device. You upload that file to your own Colab session, where it becomes the first frame of the video. It is never sent to this website.
Which models does it use?
Video: Wan2.1 VACE-1.3B through the official Wan-Video/Wan2.1 repository (your photo is the first frame). Script: Qwen2.5-1.5B-Instruct, with curated templates as a fallback. Voice: Piper TTS. Music: an original synthesised loop.
How long is the video, and how long does it take?
The ad is 15 seconds. The AI motion is a clip of about 2 seconds (33 frames) that plays forward and backward in a smooth loop with a slow zoom and pan, so it repeats; the voice, captions, music and the last 3 seconds (product name, your call to action and “WhatsApp Today”) carry the ad. Estimated time: 15–25 minutes for the first run, when about 22 GB of models download once, and 5–8 minutes for later runs. These are estimates that have not been measured on a real T4.
What if it runs out of memory?
The notebook retries with a shorter AI clip (25, then 17 frames). If the clip still cannot be made, it renders the ad with a slow zoom and pan on your photo instead and tells you so.
Why do I upload a launch file in Colab?
Colab cannot receive settings through a link. So this page writes one file with your photo and choices, and the notebook reads it in its single upload prompt. If you skip it and upload only a photo, the notebook asks for the name, language and call to action on the spot.
Can I use the videos commercially?
Check the licences first. The English voice is public domain and the Bangla voice is CC BY-SA (credit required), but the Hindi voice is licensed for non-commercial use only. Wan2.1, Qwen and the Noto fonts allow commercial use. AI-written lines are not verified facts, so review the script before publishing.