Agent-ready Figma automation
Give AI agents a structured Figma canvas API: create, edit, validate, audit, export, and compare designs from one local CLI and plugin.
$ ahd-figma batch launch-screen.json
OK 42 operations completed
$ ahd-figma command document.accessibility_audit \
'{"file":"launch-screen.json"}'
OK WCAG-oriented checks passed
$ ahd-figma command tokens.export \
'{"format":"css","variablesFile":"figma-vars.json"}'
OK wrote tokens.cssBuilt for second-pass agent iteration
The CLI gives agents concrete feedback instead of only screenshots: schemas, dry runs, linting, accessibility findings, parity scores, and exportable artifacts.
Fuller Figma surface
Document, page, node, layout, text, paint, effect, component, variable, export, REST, Dev Mode, Motion, Shader, Slots, Slides, and FigJam guards.
Tokens and variables
Export presets or live variable snapshots to JSON, CSS, Tailwind, Swift, and Android outputs with config-driven paths.
Accessibility audits
Catch contrast, line height, target size, non-text contrast, color-only communication, focus risk, and semantic naming issues.
Proof loops
Screenshot artifacts, visual verification, structured errors, local jobs, and release gates make agent work inspectable.
Agent-safe output
Use JSON, JSONL, compact fields, simple jq extraction, structured errors, and dry-run validation for reliable automation.
Local relay and plugin
The WebSocket relay and bundled Figma plugin keep canvas operations local while still exposing MCP and CLI workflows.