Wask launched its AI creative feature with each::labs as the inference layer from day one. Today it handles thousands of generations a minute at peak, with no model layer to own.
Shipped the day the each::labs partnership started.
The customer
Wask is an advertising platform for performance marketers. Its AI creative feature turns one product photo into ready-to-run ads: template ads, UGC videos, template videos and product shots that copy the look of a reference image, in every platform size.
The problem
Building template video generation in house meant two developers wiring AI providers together for 6 to 8 weeks before the first user could press generate. After launch, a bespoke inference layer would need upkeep every time new models landed, all hours taken away from Wask’s real product and roadmap.
How they use each::labs
The entire creative side of Wask runs on each::labs. Hosted models generate each creative from the template, the product image and the prompt, and the platform handles the models. Fallback is on, so if a model fails, the request moves to the next one and the creative still comes back to the user. New models on each::labs make the same flow better automatically, and the team answers fast when anything tricky comes up.
The output
One product photo in, ready-to-run creatives out. Shown with demo brands.
Product photoVideo · UGC ad
Product photoImage · Template ad
Product photoImage · Product shoot
Product photoVideo · Product film
The results
Numbers, before and after.
Time to launch the AI feature
8 weeks → Day 1
Before
After
Developers wiring AI providers
2 → 0
Before
After
Inference layers to maintain
1 → 0
Before
After
Customer story
More like a partner in AI creative.
“The entire creative side of Wask runs on each::labs. Every new model update makes our work easier, and the team responds fast when we hit anything tricky.”
Umut GülAI Expert Specialist at Wask
Ready to close the gap?
Bring the idea. Build the workflow. Produce the result.
