Every AI plugin wants $30/month. We charge $0. Bring your own Claude, Gemini, Codex, or Copilot — we’ll plug it into Figma with 88 tools for screens, tokens, accessibility, and code. No trial. No paywall. No catch.
No credit card. No signup. No account.
Opus 4▾
Claude (Opus 4)8 flagship features. 88 tools. Zero paywalls. If a feature is on this page, it ships today — not next quarter.
Type a prompt. Get a screen. iOS, Android, web — with your design system tokens, not some generic template.
See it anywhere? Rebuild it here. Drop an image or URL. Watch it land in Figma with real auto-layout — not a flat export.
Your tokens, done in one pass. Primitive → Semantic → Component — generated from whatever you’ve already built. No manual setup.
Ship accessible or ship nothing. Full WCAG 2.2 — contrast, 44px targets, focus, reading order. With fixes, not just flags.
No skipped variants. Ever. Every state, size, and combination — built to Carbon DS quality standards.
Redlines, automatic. EightShapes-style anatomy with token-linked callouts drawn on your canvas. Devs stop asking.
Design in Figma. Code in VS Code. Simultaneously. Live WebSocket bridge — build a component, get .tsx + .css in your repo.
Your design, judged by Claude. Hierarchy score, cognitive load rating, brand alignment — node by node, no emotion, no meeting.
Custom skill files that teach the AI how to perform complex design tasks with precision.
144+ variants generated
Pixel-accurate recreation
Full screens from prompts
Auto token hierarchy
WCAG 2.2 compliance
Auto redline specs
AI-powered design review
144+ variants generated
Pixel-accurate recreation
Full screens from prompts
Auto token hierarchy
WCAG 2.2 compliance
Auto redline specs
AI-powered design review
Full DS from scratch
Desktop to mobile
Consistent icon sets
Auto-connect flows
Full theme variant
Dev-ready specs
Real-time design QA
Full DS from scratch
Desktop to mobile
Consistent icon sets
Auto-connect flows
Full theme variant
Dev-ready specs
Real-time design QA
Bring your own AI subscription. We do the rest. The only plugin that plays nice with Claude, Gemini, Codex, and Copilot — all at once.
You need at least one AI provider. All four work with your existing subscription — no extra API costs.
| Provider | Install | Log in |
|---|---|---|
| Claude | claude.ai/download | claude login |
| OpenAI Codex | npm install -g @openai/codex | codex login |
| Google Gemini | npm install -g @google/gemini-cli | gemini (opens Google sign-in) |
| GitHub Copilot | npm install -g @github/copilot | copilot then type /login (GitHub device flow) |
Open your terminal and run this one command:
npx @sarjallab09/figma-intelligence setup
The setup wizard will automatically:
Figma Intelligence)| Command | What it does |
|---|---|
npx @sarjallab09/figma-intelligence start | Start the relay server |
npx @sarjallab09/figma-intelligence stop | Stop the relay server |
npx @sarjallab09/figma-intelligence status | Show relay + provider status |
npx @sarjallab09/figma-intelligence doctor | Diagnose any issues |
npx @sarjallab09/figma-intelligence@latest update | Update plugin to latest version |
npx @sarjallab09/figma-intelligence start before opening the plugin.The setup command already copied the plugin path to your clipboard and opened the plugin folder in Finder. Just follow along:
The plugin import option only shows up when a design file is open.
Press Shift + I or click the grid icon in the top toolbar.
Click the Plugins tab at the top → click the ··· menu in the top-right corner of the panel → select “Import plugin from manifest…”
manifest.jsonTwo ways, pick whichever is easier:
The path is already in your clipboard.
Browse to the plugin folder:
That’s it for Step 3. Head to Step 4 below to launch it.
You can open Figma Intelligence any time using either of these:
Type what you want in the chat box at the bottom of the plugin panel:
"Create a login screen with email and password fields" "Add a navigation bar with 4 menu items" "Design a card component with a photo, title, and a blue button" // Attach a screenshot, then: "Recreate this layout in Figma"
To attach an image: click the paperclip icon next to the chat box.
| Tab | What it does |
|---|---|
| Chat | Ask questions — the AI answers but does not change your Figma file |
| Code | The AI builds and edits your Figma design directly using MCP tools |
| Design + Code | Same as Code, but also generates React / Vue / Svelte source code and writes it to your VS Code workspace |
The most powerful mode. From a single prompt, the AI:
src/components/ automaticallyIf VS Code is not running, code output still appears in the Figma chat as text.
Use the provider badge in the plugin header to switch between Claude, OpenAI, Gemini, and Copilot at any time. Setup registers the same MCP server for all providers, so switching reuses the running relay automatically.
| Phase | Tools | Examples |
|---|---|---|
| Visual Intelligence | 7 | Screen cloning, visual audit, accessibility audit, sketch-to-design |
| Design System Accuracy | 6 | Intent translation, layout intelligence, variant expansion, linting |
| Generation & Scaffolding | 14 | Page architect, swarm build, prototype wiring, composition builder |
| Sync & Bidirectionality | 9 | Token export (16 formats), component code gen, handoff specs |
| Governance & Health | 15 | Health reports, token docs, design system scaffolding, DTCG validation |
| Node Operations | 38 | Create, read, update, delete nodes, variables, styles, components |
| Problem | Fix |
|---|---|
| “npm: command not found” | Node.js is not installed — download from nodejs.org |
| Plugin shows “Server offline” | Run npx @sarjallab09/figma-intelligence start in your terminal |
| Plugin not found in Figma | Import it: Shift+I → Plugins tab → ··· menu → “Import plugin from manifest…” → Documents → Figma Intelligence → manifest.json |
| Can’t find the plugin folder | It’s in ~/Documents/Figma Intelligence/. On Mac, use Cmd+Shift+G in the file picker and paste the path. |
| Chat does nothing / no response | Re-authenticate your AI provider: claude login / codex login / gemini |
| Gemini doesn’t build in Figma | Run npm i -g @google/gemini-cli, then gemini to sign in, then re-run setup |
| Wrong or expired Figma token | Re-run npx @sarjallab09/figma-intelligence setup and paste your new token |
| Port 9001 already in use | Run npx @sarjallab09/figma-intelligence stop then start again |
| Not sure what’s broken | Run npx @sarjallab09/figma-intelligence doctor for full diagnostics |
npx @sarjallab09/figma-intelligence@latest updateThis copies the latest plugin files to your local folder. Just close and reopen the plugin in Figma — changes are picked up automatically.

Nilmani is a product designer obsessed with closing the gap between design and development. He built Figma Intelligence to eliminate the copy-paste loop between AI tools and Figma — enabling designers to generate tokens, clone screens, audit accessibility, and ship code without ever leaving their canvas.
Figma Intelligence started as an experiment to see how far AI could go inside a design tool. It grew into a full plugin system connecting multiple AIs, 88 tools, and a live VS Code bridge — all from a single chat panel inside Figma.

Ram is a Product Designer with 13+ years of experience leading and building multi-disciplinary design teams.
Ram is deeply involved in AI research and building AI-powered design tools. He co-created Figma Intelligence to bring AI directly into the design workflow — exploring how large language models can generate production-ready components, audit accessibility, and bridge the gap between design and code. His work focuses on making AI usable for designers, not just engineers.