Prompt workspace for AI image & video

Stop rewriting prompts. Build them from parts.

Every library here shows you the move before you use it. Pick a camera movement, a performance and a look, then compile the whole thing into the syntax your model actually reads.

Compiling for Seedance 2.5
A lone soldier scrambles up a muddy trench wall as shells land behind him
CameraLow tracking
PerfDread
LookBleach bypass
FramingMS · low angle · 35mm

Live now

The libraries

These already exist as standalone tools. They now live in one place, share one design, and will feed the same builder.

Roadmap

What comes after launch

Shipping order, not a wishlist. Each one earns the next.

01
Reverse engineerDrop in an image or a clip and get the prompt structure back, split into the same parts the libraries use.
Next
02
Prompt saverYour own prompts, organised by project. The first thing here that belongs to you and not to us.
After that
03
Character sheetsLock a character once — appearance, wardrobe, angles — and bind it into every shot in a scene.
Q4
04
MCP serverYour saved presets available inside Claude and ChatGPT, so the prompt gets built where you already work.
Q4
05
Creator profilesPublish your style as a profile others can follow, subscribe to and build on.
Later