Menu

I built a tool that recreates the Bicep Heart trend from one photo - week one numbers inside
📰

I built a tool that recreates the Bicep Heart trend from one photo - week one numbers inside

DEV Community: machinelearning·Steve Ai Pro·4 days ago
#gWZZ1jmp
Reading 0:00
15s threshold

The Bicep Heart trend is a short-form video pose that went viral: two people standing back to back, flexing, hands forming a heart behind their heads. Every version has the same pose, the same rhythm, the same camera framing.

I kept seeing people ask how to make one without filming. Text-to-video tools can sort of do it, but every render invents new motion, so the output rarely looks like the trend. So I built Bicep Heart : you upload one photo, and it renders a 7-second vertical video of you doing the exact routine from a fixed reference clip.

How it works

One JPEG/PNG/WEBP photo of a person, up to 10 MB

Motion comes from a fixed reference clip via a motion-control video model (Kling through APIMart, roughly $0.40 per render), so pose, timing and camera framing stay identical across outputs

Google sign-in, English and Chinese UI

Credits are one-time packs that never expire; failed renders refund automatically

Stack

Layer Choice

Runtime Cloudflare Workers + TanStack Start SSR

Data D1 for…

Continue reading — create a free account

Join HashtagPLUS to read full articles, follow hashtags, vote, and join the conversation.