Connect your repo or project folder, design on its real screens, and push — Lindi Code or your agent writes it into the code.
Connect a repository or a folder on your machine, and Lindi puts the app’s screens on the canvas.
Auto layout behaves like real layout, and variables are the values your code reads — nothing to translate later.
A push writes the design change into your code — by Lindi Code, or by the coding agent you already use.
Describe what you need. Lindi Code plans it, waits for your approval, and builds it — screens and code.
Claude Code, Codex, Cursor — any agent that speaks MCP reads the design as data and builds with you.
Measurements, annotations, variables and code for any selection — and a list of every screen ready to build.
The screens come from your code, so what you design is the app — with frames, auto layout, components and variables, and your team on the canvas beside you.
In Build mode, say what you need in plain words. Lindi Code asks what it needs to know, shows you a plan, and builds it once you approve — straight into your code.
Push from the canvas and the change is written into your code — by Lindi Code, or by Claude Code, Codex or Cursor. Your team reads it and ships it the way it ships everything.
A designer changes a real screen and pushes, Lindi Code writes it into the code, and the team ships it the way it always does. Nobody redraws, respecifies or retypes anything.
Designing on the canvas never touches your code. Only a push does, and you choose when.
A push lands as ordinary changes in your project, so your usual review covers it before anything ships.
An owner can switch AI off for the whole company, and every change to that setting is recorded.
Lindi is free for its early users. Tell us about your team and we answer within 48 hours — yes or no.