I’m building a stable prototype with Figma Make, and the credit model is a practical blocker.
-
I’m already at ~500 prompts/executions on this project.
-
Figma’s own positioning puts 3,000 credits/month at ~50–70 Make prompts (~43–60 credits per prompt on average).
-
If my workflow needs ~500 prompts, that’s ~21,500–30,000 credits, or ~7–10 months of the included allowance.
-
I’ve also seen 4,250 credits disappear without reaching a usable prototype outcome.
The main problem is that because Make is agentic, the cost isn't fixed per prompt and you can't accurately predict it before running. What the agent chooses to do (actions, files, iterations, model) determines billing. Because you can't confidently plan or iterate, it's a budget black box.
Make only seems feasible for small demos at the moment, or if you manage to iterate with almost no reruns. Sometimes it is not enough even for a small prototype or creative exploration.