The first screen appeared quickly, but can you afford a few more changes?
Describe the flow you want in Figma Make, and screens and interactions appear quickly. Instead of drawing a layout piece by piece on a blank canvas, you can describe screens, states, and behaviors in natural language and review a working UX prototype first.
The challenge starts after you see the first result. As you move buttons, add missing states, and align the result with brand components, credits keep dropping. Since usage varies by prompt, it can be hard even to predict how far your remaining credits will take your revisions.
The practical value of Figma Make is less about a promise to finish in a few hours and more about narrowing the flow to validate, creating the first draft quickly, and moving to editing after checking actual usage.
64 credits is one user’s example, not a fixed rate
Figma Make does not have a fixed price such as “one edit costs this many credits.” According to the official documentation, usage varies based on the selected model, task complexity, the number of tasks the model performs, attached files and libraries, and accumulated conversation context.
For the default model as of February 2026, Figma’s published estimates are 30+ credits to change a font, 75+ credits to add interactions to an attached design, and 100+ credits to create a new app. These are examples only, and actual usage may vary.
| Published example | Usage | Interpretation |
|---|---|---|
| Font change | About 30+ credits | Official estimate for the default model in February 2026 |
| Making an attached design interactive | About 75+ credits | An official estimate that varies with attached context |
| Creating a new app | About 100+ credits | An official estimate that varies by scope and complexity |
| Moving a toast message by 50px | 64 credits | A forum report with undisclosed reproduction conditions |
| Revision and error-checking prompts | 600+ credits | An individual example posted by another user |
The 64-credit and 600-credit examples were posted on the Figma forum, but the model used, full conversation history, and attached context were not disclosed. They therefore cannot be generalized as Make’s average cost or standard rate. Still, they are useful warnings that even seemingly small changes can consume more credits than expected depending on context.
As of August 25, 2026, the monthly included credit amounts for a Full seat are 500 for Starter, 3,000 for Professional, 3,500 for Organization, and 4,250 for Enterprise. They reset every month, and unused credits cannot be carried over or transferred to another user. Cash charges and included-credit usage are not the same thing, so also check whether an administrator has configured additional subscriptions or pay-as-you-go billing.
Design-system and editability concerns have improved
The view that Make results cannot be brought in as editable layers or reflect an existing design system is now outdated. However, the availability of related features differs by plan and seat.
With Make kits, you can provide Make with code components from npm packages or styles, tokens, and usage guidance from Figma libraries. Make kits are available to Full seats on paid plans. Project materials such as PDFs, CSVs, JSON files, and screenshots can be attached to prompts.
Copy design is a paid-plan feature that imports a snapshot of a Make preview into a Figma Design canvas as structured, editable layers. You can then adjust details such as spacing and typography on the canvas.
Editable layers do not mean a perfect mapping to your company components or finished production code. After import, you should separately review the layer structure, component connections, responsive states, accessibility, and error and empty states. Make kit results also depend on the quality of the prepared tokens and usage guidance.
Check account requirements before you begin
For a first test, you need a Figma account, available AI credits, and one user flow to validate. Also check your current plan and seat, and whether AI features are enabled for the organization first so you do not get blocked later. In Organization and Enterprise, an administrator must enable AI features including Figma Make, and generation and editing permissions may also differ by seat type.
- You need a Full seat on a paid plan to connect a Make kit.
- You need a paid plan to copy results into design layers with Copy design.
- If you cannot use both features, define success for this test as validating the core flow in the Make preview and recording actual credit usage.
To check remaining credits, open an editable Figma file and select AI balance from the main menu. Note the reset date as well.
For your first test, define completion criteria before screen count
The most realistic way to estimate credit use is to generate a small version of one real team flow. Once you have checked your balance and account conditions, begin in this order.
- Create a new Figma Make file.
Return to the Figma file browser, open Drafts, and select Make in the upper-right corner. Once the new Make file opens, you can enter a prompt in AI chat. - Choose one path to validate.
Rather than the entire product, choose a flow with a clear beginning and end, such as “email sign-up → choose interest categories → arrive at Home.” The success criterion is whether users can click through the core path and verify required states. - Connect only context your account supports.
If you have a Full seat on a paid plan, connect the necessary Make kit. Otherwise, attach relevant starting frames or precisely scoped data, documents, and screenshots. Large unrelated files and ambiguous materials can increase the context the model must process, increasing usage and output variability. - Write the first AI chat prompt as a short brief.
Include the goal, user, screens and states, key actions, constraints, and completion criteria at once. For example: “Create a mobile prototype where a new user signs up by email, selects three interest categories, and arrives at Home. Include only the three screens for sign-up, category selection, and completion, and do not let the user proceed until required selections are made. Exclude payments and a real backend.” This closes the scope. - Limit follow-up requests to one change.
State what to change, how to change it, and what to preserve, such as: “Only fix the Next button on the category-selection screen to the bottom; keep the colors, copy, and rest of the layout.” The official recommendation is not to combine unrelated edits in one request. - Once it works, record usage and finish according to your account.
Write down the cumulative credits and number of revisions when the core path first passes. On paid plans, use Copy design to bring it to the canvas and review its structure, components, accessibility, and states. For accounts without Copy design, finish the test by confirming the core flow and required states in the preview and retaining the credit record.
For team-adoption decisions, separate “first generation” from a “usable draft”
Do not compare only the time from pressing Generate to seeing the first screen. Look at the time and credits needed to reach a draft that passes completion criteria, and, for supported accounts, the human cleanup time after moving it to the canvas.
What to record from one test
- Plan, seat, and whether organizational AI is enabled
- Selected model and start and end times
- Cumulative credits used for the first generation and revisions
- Number of follow-up prompts and reasons for regenerating
- Whether the core path and required states passed
- If you used Copy design, time spent replacing components and fixing accessibility
If the first generation is fast but revision and cleanup take a long time, start by improving the prompt scope or Make kit. Conversely, if you can validate the core flow with few revisions and finish canvas cleanup quickly, you can expand use to similar early-exploration tasks.
Figma Make is closer to a tool that turns ideas into working drafts for earlier validation than one that automatically delivers finished screens. Its speed shows up in the first generation, but its real efficiency only becomes clear when you record credits, revision count, and design-system cleanup.
If you want to go deeper
Figma Make FAQs explains how to start a Make file from Drafts, permissions by seat, organizational AI enablement requirements, and Copy design availability. help.figma.com
How AI credits work covers included amounts by plan, factors that affect usage, and official estimates. help.figma.com
7 tips for using Figma Make credits more efficiently provides official guidance for scoping your first prompt and follow-up revision requests. figma.com
Get started with Make kits explains eligible plans and seats for Make kits and the design-system materials you can include. help.figma.com
