Figma logo

Figma

OrganizationPopular
heygen-com
figma

Bring Figma into a HyperFrames video - rendered assets, brand tokens, components, storyboard frames as reconstructed motion (frames read as states, not slides), shaders and connector-assisted motion. Use when the person pastes a figma.com link or asks to use a Figma design, frame, logo or brand.

Overview

Publisherheygen-com
Repositoryhyperframes
Skill namefigma
Stars
73.3K
Forks
6.6K
Bundled files
1
LicenseApache-2.0
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 1 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by heygen-com on GitHub. Read the source before you install it.

Installation

Install the Figma AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/heygen-com/hyperframes/tree/main/skills/figma
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/heygen-com/hyperframes.git /tmp/hyperframes
mkdir -p .claude/skills
cp -r /tmp/hyperframes/skills/figma .claude/skills/figma
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Figma in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Figma on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Figma is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Plugin installs: Before setup or freshness commands, follow plugin execution rules when this skill is inside a HyperFrames plugin.

Figma → HyperFrames

Every import happens at build time and freezes a local file with its source recorded under .media/, so renders never call Figma and re-running a command re-imports only what changed. Assets, tokens and components go over REST with a token; motion and shader data need a Figma connector, or a native export from the person.

Access

FIGMA_TOKEN in the shell or the project .env. Check it exists before the first command; if not, tell the person how to make one and stop: figma.com → Settings → Security → Personal access tokens, read-only File content and File metadata, plus Library content for tokens on non-Enterprise plans. They set it themselves; never ask them to paste it into the chat.

ErrorMeaning
BAD_TOKEN (401)Expired or revoked: make a new one
FORBIDDEN (403)The message names the missing scope, or the file isn't visible to the account
REQUIRES_ENTERPRISE (403, variables)Not a failure: the published-styles fallback already ran
RATE_LIMITED (429)Already retried with backoff; wait a minute or import fewer nodes per call

Commands

  • A layer, logo or image: npx hyperframes figma asset '<url or fileKey:nodeId>' [more refs] --format svg|png --scale 2 --description "<what it is>" [--entity "<brand mark>"]. SVG for vectors and logos, PNG at 2x for raster. Several nodes of one file in one call render in a single request; --entity lets /media-use find a brand mark later.
  • Brand colours and type: npx hyperframes figma tokens <fileKey>. Run it before component and paste the printed entries into data-composition-variables. Without Enterprise it falls back to published styles. A connector can supply variables there instead: cache its raw response in .media/figma-cache/ and write the bindings to .media/figma-bindings.jsonl ({kind:"binding", figmaId, sourceFileKey, compositionVariableId:"figma:<name>", version}).
  • A frame as editable HTML at exact geometry: npx hyperframes figma component '<url or fileKey:nodeId>', packaged under compositions/components/<name>/. A fill bound to an imported token comes out as var(--slug, #literal), so one variable re-brands every import; an unknown token stays a literal marked data-figma-unresolved: import that library's tokens and re-import. Ask once which file a library is; never match by hex.
  • Check hero components against Figma's pixels (figma asset <same node> --format png). Text drift is the usual cause; report drift you can't fix instead of hand-tweaking the import.

Motion (connector)

Usage beacon: these phases have no CLI touchpoint, so run npx hyperframes events --skill=figma-motion when you begin and npx hyperframes events --skill=figma-motion --event=skill_completed --outcome=success|error when done. Same for shaders (figma-shaders) and storyboards (figma-storyboard).

  1. Fetch motion context for the parent frame in one recursive request; cache the raw JSON in .media/figma-cache/.
  2. Normalise with motionContextToDocs(raw, { selectorFor, repeat }) from @hyperframes/core/figma, never by transcribing keyframes. It reads the motion.dev snippets (the CSS snippets stretch durations), drops loop-wrap tail keyframes and keeps bezier eases. selectorFor returns the ids from the component import.
  3. motionToGsap(doc) → emitTimelineScript(spec), injected after the GSAP and CustomEase tags: paused, finite, a literal window.__timelines key. Named eases outside the mapped set fall back to linear; say so.
  4. Verify against Figma's own export: node <SKILL_DIR>/scripts/verify-motion.mjs --reference <export.mp4> --render <render.mp4> --crop WxH+X+Y. It compares motion between samples, so static drift cancels out, and fails under 15 dB (faithful ≈ 20+, diverging ≈ 5). Measure --crop from the render's card edges. A fail means fix the translation, not the threshold.
  5. A track that won't translate: export it, freeze the MP4 and place it as <video class="clip">.

Shaders

Figma's export path flattens shaders to their base colour, and shader source is only reachable for library-published styles. Ask the person for a native export of that frame (PNG or MP4) and place it as an asset or clip.

Storyboards: frames are KEYFRAMES, not slides

Two frames holding the same element describe that element's state through time: animate the element between the states. Playing the frames back to back as stills is the failure mode.

  1. Scenes are every frame-sized node in the section, named frames and loose full-frame rectangles alike: filter by size, not type or name. Order by absoluteBoundingBox.x.
  2. Match children across neighbouring frames by name first, then by similar size and position. Tween geometry between states; crossfade two exports in place when the pixels changed. Export one asset per element, not one still per frame. Unmatched children enter or exit at their scene's beat.
  3. Text nodes below the strip are motion notes for the scene above them, not on-screen copy. A note about motion inside a scene ("lines reveal one after another") means import that frame as a component and animate it.
  4. Export in batches of about 4 scene frames per call: past about 12 ids /v1/images times out.
  5. Frames that are one product UI in successive states: rebuild the UI as live DOM: code what changes state, freeze what doesn't, and treat each frame change as an interaction to perform (the cursor moves, clicks, the screen responds), not a tween to apply.
  6. Stills are the fallback, only for frames with no shared elements.

Never leave a Figma URL in the composition.

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Figma AI skill do?

Bring Figma into a HyperFrames video - rendered assets, brand tokens, components, storyboard frames as reconstructed motion (frames read as states, not slides), shaders and connector-assisted motion. Use when the person pastes a figma.com link or asks to use a Figma design, frame, logo or brand.

Why use Figma on TypingMind?

Because you install it once and use it with any model. Figma is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Figma in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/heygen-com/hyperframes/tree/main/skills/figma. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Figma?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Figma?

As many as you like. As long as a model supports skills, you can use Figma with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Figma AI skill free?

Yes. It is published on GitHub by heygen-com under the Apache-2.0 license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇