Figma Make
Also known as: Figma, Inc., Make by Figma, figma.com/make
Figma Make is Figma's prompt to app builder that turns ideas and existing Figma designs into code backed, visually editable prototypes and web apps, with plan mode, a Payload or Supabase backend, Make kits that carry a team's design system, and MCP connectors to tools such as Notion, Atlassian and Linear. Builders choose between GPT, Claude and Gemini models, publish to a figma.site address or a custom domain, and Enterprise adds data residency.
Figma Make is Figma's prompt to app builder, sold inside Figma's plans alongside Figma Design, Dev Mode, FigJam and Sites, that turns ideas and existing Figma designs into functional prototypes, web apps and interactive UI. Everything it builds is code backed and visually editable: previews can be adjusted with a Figma style properties panel, marked up with annotations, copied back to the canvas as design layers, or changed directly in the code editor.
In the chat, plan mode drafts an editable plan before building, and Make then generates the app, including a Payload backend with a database, file storage, authentication and an admin panel, or a connection to the team's Supabase project.
Make kits carry a team's npm packages, library styles and guidelines into every build, attachments and MCP connectors to Notion, Atlassian, Linear, GitHub and others bring in specs and data, web search and fetch ground it in public pages, and a model selector switches between the default and GPT, Claude and Gemini models.
Apps publish to a figma.site address or a custom domain, and Make on a local codebase, which edits a team's real GitHub repository and opens pull requests, is in closed beta.
Figma Make is included on Full seats. The Starter plan is free with 500 AI credits a month, at most 150 a day. Professional costs $20 a month per Full seat, or $16 billed annually, with 3,000 AI credits; Organization is $55 and Enterprise $90 per Full seat a month billed annually, with 3,500 and 4,250 credits. Dev and Collab seats can try Make in drafts, and admins can add AI credits by subscription or pay as you go.
Figma Make fits product teams, designers and PMs who already work in Figma and want realistic, on brand prototypes and web apps built from their own design system, specs and data, with enterprise compliance and data residency. It fits less well for teams that need an agent working in their production codebase today, running tests, or driven by an API from outside Figma.
Vendor details
Canonical URL
https://www.figma.com/make/
Category
Coding agent
Subcategory
Prompt to app builder inside Figma that turns designs and prompts into code backed, visually editable prototypes and web apps with backends, design system kits and MCP connectors
Funding status
Product of Figma, Inc., listed on the New York Stock Exchange as FIG since its 2025 initial public offering.
Company status
first party product
Use cases & customers
Primary use cases
Target customers
Deployment options
Integrations
MCP client with 21 verified connectors (Amplitude, Asana, Askable, Atlassian, Box, Contentsquare, Dovetail, Granola, GitHub, Hex, Linear, Mapbox, Marvin, Mobbin, monday.com, Notion, Slack, Sprig, Twilio, Zapier, zeroheight) plus custom MCP connectors on paid plans; Payload backend scaffolding; Supabase connection; npm packages and Figma libraries through Make kits; web search and fetch; Figma MCP server at mcp.figma.com/mcp exposes Make code resources to outside agents.
In practice
A product designer attaches a checkout flow from Figma Design, uses plan mode to agree the structure, and Make builds an interactive prototype in the team's Make kit that stakeholders try in a full screen preview and comment on.
A PM connects Notion and Linear, asks Make to build a first draft prototype from the PRD and open tickets, then reruns the connector when the spec changes and lets Make update the app.
A marketing team asks Make for a restaurant website with an editable menu, and Make scaffolds a Payload backend with collections and an admin panel, seeds sample items and publishes the site to a custom domain.
Sources & related URLs
Related / legacy domains
Research sources
Agentic Index coverage score
11.5 / 14 capabilities · 82%
| Integrations & Tool Calling | Full |
|---|---|
|
Make works as an MCP client with 21 verified connectors, among them Notion, Asana, Atlassian, GitHub, Linear, Slack, Box, Amplitude, Hex, monday.com, Twilio and Zapier, each signed in through the outside tool's OAuth, and write tools such as creating Jira issues or Notion pages can be turned on per connector. Paid plans add custom MCP connectors, which organizations can publish for their members, and an app can be wired to a Supabase project or a Payload backend that holds private API keys and secrets. npm packages bring a team's component libraries into the generated code. Sourcehelp.figma.com/hc/en-us/articles/35440096186007-use-verified-partner-mcp-connectors-with-the-figma-agent-and-figma-makeread 2026-10-18 |
|
| Workflow Orchestration | Full |
|
Plan mode reads the project, its code and any attached designs, writes a structured plan to plan.md that can be edited or revised in chat as often as needed, and executes it when the builder clicks Build. Asked for an app that stores data, Make designs the data model, creates the Payload collections, seeds sample records and builds the interface with sign in and an admin panel. More than one chat can run at once in a single Make file, so a new change can start while another is still generating. Sourcehelp.figma.com/hc/en-us/articles/40830441709719-use-plan-mode-in-figma-makeread 2026-10-18 |
|
| Knowledge Grounding & RAG | Full |
|
Prompts can carry attached Figma frames and components, images, videos, audio, text files and PDFs, plus Community content. Make kits sync a team's npm packages, library variables and styles and written guidelines so generated code uses the real design system, and a guidelines.md file sets rules for each project. Connectors pull PRDs, tickets and research from Notion, Confluence, Asana, Linear, Dovetail and others and can be rerun when the source changes, while web search with citations and web fetch bring in public pages. Sourcehelp.figma.com/hc/en-us/articles/39241689698839-get-started-with-make-kitsread 2026-10-18 |
|
| Human Oversight & Guardrails | Full |
|
Plan mode holds generation until a person approves the plan with Build, and when a prompt needs a connector, Make names the connector and tool it will call and waits for Run, with write tools enabled per connector in its settings. Admins can turn AI, connectors and website publishing off for a team or organization, the Review tab shows each change as additions and deletions, and code edits create checkpoints that revert to the last conversation step. A published app changes only when someone pushes an update live. Sourcehelp.figma.com/hc/en-us/articles/40830441709719-use-plan-mode-in-figma-makeread 2026-10-18 |
|
| Security, Identity & Governance | Full |
|
Figma holds SOC 2 Type 2 and SOC 3 reports, ISO 27001, 27017, 27018 and 27701 certification, ISO 42001 for its AI management system, FedRAMP authorization, C5 and TISAX. Enterprise adds SCIM seat management and Governance+ controls for access, sharing and compliance, and admins decide whether AI features and content training are on for their team or organization. Verified connectors are vetted against Figma's security and privacy standards, and shared Makes can be protected with a password. Sourcefigma.com/securityread 2026-10-18 |
|
| Observability & Auditability | Full |
|
Version history records every AI and manual edit in a Make, the Review tab shows the code each prompt changed, and hovering under a finished prompt shows the AI credits it consumed. Admins see credit use across the plan in an AI credits dashboard, with people at their limits and reset dates, and Enterprise organizations get an AI Usage API. Published apps carry no built in analytics, though a Google Analytics ID or other snippets can be added in settings. Sourcehelp.figma.com/hc/en-us/articles/33459875669015-how-figma-ai-credits-workread 2026-10-18 |
|
| Memory & State Persistence | Partial |
|
Make kits, guidelines and custom skills persist outside any single chat: a kit can be attached to new Makes across a team or organization, and skills package frequent workflows as reusable instructions. Version history keeps every AI and manual edit so earlier states can be previewed, favorited or restored, and conversations stay in the file for collaborators who open it. Make keeps no learned memory of a person's preferences beyond those saved kits, guidelines and skills. Sourcehelp.figma.com/hc/en-us/articles/40283639496599-create-and-manage-custom-skills-in-figmaread 2026-10-18 |
|
| Deployment & Data Residency | Full |
|
Enterprise organizations can keep Figma Make files at rest in the EU, Australia, India or Brazil through Figma's data residency option on AWS, though anything published to the web falls outside it. Apps publish to a figma.site address or a custom domain, on AWS with Cloudflare routing, and organization admins can turn publishing off. Code can be downloaded as a zip, and Make on a local codebase, which opens a GitHub repository and pushes branches and pull requests, is still a closed beta. Sourcehelp.figma.com/hc/en-us/articles/15643274574871-enable-localized-file-hostingread 2026-10-18 |
|
| Prebuilt Agents / Templates / Packs | Full |
|
Templates turn a polished Make into a starting point that teammates across a team or organization can use under the guidelines its author sets, and the Figma Community holds Makes to remix along with skills that can be installed. Make kits package a design system as reusable AI context, organization admins can manage kits for everyone, and example prompts such as an onboarding flow, a data dashboard or a gradient gallery start a new Make in one click. Sourcehelp.figma.com/hc/en-us/articles/34716344138519-create-and-update-a-template-in-figma-makeread 2026-10-18 |
|
| Triggers & Channel Coverage | Partial |
|
Work begins in the chat of a Make file in the browser or desktop app, typed or dictated by voice, with frames pasted in from Figma, and annotations or markups drawn on the preview can be sent together as one prompt. Makes can also be opened and built on a phone, where they run natively and comments can be answered in context. Nothing starts Make on a schedule, from a chat channel or from an outside event; each change begins with a person in the file. Sourcehelp.figma.com/hc/en-us/articles/31304412302231-explore-figma-makeread 2026-10-18 |
|
| Model Flexibility & Routing | Full |
|
A model selector in the chat lets the builder switch at any time between the default model and GPT, Claude and Gemini models from OpenAI, Anthropic and Google, and each file reopens with the last model chosen. Each model has its own credit pricing, and some features depend on the choice: plan mode and web search work with the default and Claude Opus models but not on Gemini. The default model changes over time as Figma updates it. Sourcehelp.figma.com/hc/en-us/articles/36400680326551-select-an-ai-model-to-use-in-figma-makeread 2026-10-18 |
|
| APIs / SDKs / MCP Extensibility | Partial |
|
The Figma MCP server, remote at mcp.figma.com/mcp or local through the desktop app, lets outside coding agents gather code resources from Make files as context for production work, alongside design context from Figma Design and FigJam and writing back to the canvas, with Figma skills guiding those workflows. No API or MCP tool creates a Make, prompts it or puts it live from outside, so building stays in the Make chat. Sourcehelp.figma.com/hc/en-us/articles/32132100833559-guide-to-the-figma-mcp-serverread 2026-10-18 |
|
| Testing, Debugging & Optimization | Partial |
|
Every Make can be tried in a full screen preview or on a phone and adjusted with the point and edit tool, and comments collect teammates' feedback right on the prototype. Connectors to Sprig and Amplitude let teams run studies and track design variants against real users, and a Make kit is checked by prompting a test build and looking for drift from the design system. Make does not write or run tests on the app it builds, and fixes go back through the chat. Sourcehelp.figma.com/hc/en-us/articles/31304548273047-preview-a-functional-prototype-or-web-appread 2026-10-18 |
|
| Browser / Computer-use | Partial |
|
Web fetch retrieves the full content of a public page, such as a documentation site or brand guidelines, and web search finds current examples, so Make can apply what it reads to the app; pages behind a login or anti scraping protection are out of reach. That reading of public pages is the extent of it: no agent clicks through sites, fills in forms or operates a desktop, and the Browser tab previews only the app being built. Sourcehelp.figma.com/hc/en-us/articles/39715554287255-search-the-web-with-the-figma-agent-and-figma-makeread 2026-10-18 |
|
The Agentic Index coverage score grades every vendor Full, Partial or Not documented against the same 14 buyer facing capabilities, from public evidence only. Each capability links to how all vendors in the index score on it. How this evidence is graded
Pricing
Professional Full seat from $20 a month ($16 billed annually); Organization $55 and Enterprise $90 per Full seat a month billed annually
Per seat subscription by plan and seat type, monthly or annual on Professional and annual on Organization and Enterprise, with monthly AI credits included per seat and shared add on credits
Included quota
Starter 500 AI credits a month (150 a day). Professional Full seat 3,000; Organization Full seat 3,500; Enterprise Full seat 4,250; Dev and Collab seats 500. Credits reset monthly and do not roll over.
What is public
Every plan and seat price with included AI credits, monthly and annual for Professional and annual for Organization and Enterprise.
Billing mechanics
Per seat subscription; Professional monthly or annual, Organization and Enterprise annual; AI credits added by subscription or pay as you go.
Cost watchouts
Make needs a Full seat for regular use; non default models can use substantially more credits per task; plan mode adds a planning charge before the build; Governance+ is a separate Enterprise add on.
Variable cost rationale
Make prompts draw on shared AI credits whose cost per task varies with the model, the task and the context supplied, and plan mode spends credits on top of the build, so heavy prototyping pushes teams into credit subscriptions or pay as you go.
Additional watchouts
Starter users can publish to the public web only if they also publish to the Figma Community.
Overage / add-ons
When seat credits run out, paid AI features stop until reset unless an admin adds credits through an AI credits subscription or pay as you go billing; admins can set custom limits per person.
Sales call required
Mixed (some tiers require a call)
Free / trial
Starter plan: free with 150 AI credits a day, up to 500 a month; Dev, Collab and View seats can try Make in drafts
Lowest paid plan
Professional Full seat at $20 a month, or $16 a month billed annually, with 3,000 AI credits a month and Figma Make, Figma Design, Dev Mode MCP Server and team libraries
Commercial notes
Hosting for published Make apps is included in paid plans while publishing is in beta; Education plans include one published Make file.
Key ambiguities
Prices for add on AI credits and the credit cost of each model in Make are not given on the pricing page.
Missing data
AI credit add on prices and per model credit rates.
Related vendors
- Cognition — Maker of Devin, an autonomous AI software engineer
- 10Web — Agentic website platform whose specialized AI agents build, host,…
- AgentUI — Managed AI app builder for operations teams that generates and hosts…
- Aider — Open source, model agnostic terminal coding agent that edits your…
- Anthropic Claude Code — Anthropic's agentic coding system across terminal, desktop, IDE, web…
- Anything — Anything is an AI app builder whose agent turns a plain language…