Edit by hand.
Select, restyle, rearrange — on the real product, not a drawing of it. The inspector is real CSS.
- Real components, real tokens
- Auto layout that is actual flexbox
- Multiplayer cursors on the canvas
The canvas edits the real frontend. Every change lands in your repo as reviewable code.
A design tool draws a picture of the product. Engineering rebuilds the picture. From then on there are two of everything, and only one of them ships.
A design file is a picture of the product. The product moves on with every release. The picture does not, and nobody is paid to keep the two honest.
Every change is made twice — once on the canvas, once in code. Two people, two tools, and an afternoon of “is this what you meant?”
Ask who changed the button and why. The design file does not know. Git does — but only for the half of the change that made it to code.
Lindi opens your shipped frontend on a canvas. Move a button, and the button moves in the code. Every change leaves as a pull request your engineers review like any other — and the canvas shows what was merged.
Select, restyle, rearrange — on the real product, not a drawing of it. The inspector is real CSS.
Tell the agent what should change. It edits the same file and shows you the result before anything leaves.
Open the whole codebase — front end and back end — beside the canvas. What you type renders. What you draw compiles.
<article className="dish-cell">- <Thumb className="rounded-md" />+ <Thumb className="rounded-xl shadow-sm" />- <Price className="text-sm" />+ <Price className="text-base font-semibold" /> </article>
Not an export. Not a spec. A pull request, opened by the person who made the change, with the change in it.
Each plan is a capability, not a quota. You pay when other people edit with you, and again when the agent does.
Lindi is pre-release. Get in line — we count sign-ups in order and we do not invent numbers.