Read the brand. Anywhere.

Endpoints

1. GET/brand/ideal-state/spec.jsonapplication/json

   Canonical structured brand spec. Tokens, voice, audience, logo, typography, color, applications. AI-readable.

2. GET/brand/ideal-state/spec.mdtext/markdown

   Agent-optimized prose + structure brand spec. The same content as spec.json plus the human-authored BRAND_BOOK.md. Drop this into a system prompt.

3. GET/brand/ideal-state/tokens.csstext/css

   CSS custom properties for color, typography, spacing, radius, motion. Ready to import into any external site.

4. GET/brand/ideal-state/tokens.figma.jsonapplication/json

   Figma Variables payload. Push into a Figma file via the Variables API or the Figma plugin.

5. GET/brand/ideal-state/llms.txttext/plain

   Discovery file for AI tools. Points crawlers and agents to the canonical spec endpoints for this brand.

6. POST/api/brand-spec/ideal-state/demoapplication/json

   Run a named prompt from the brand spec's promptKit against Claude. Body: { promptId, variables }. Rate-limited per IP.

Use with Claude

# Curl the spec.md and pipe it into Claude as system context.
curl -s https://wrkhrs.co/brand/ideal-state/spec.md \
  | claude --system - "Write a launch email about our new pricing tier."

Use with any project

<!-- Drop the brand's tokens into any external site. -->
<link rel="stylesheet" href="https://wrkhrs.co/brand/ideal-state/tokens.css" />

Embed a brand card

<!-- Live brand-book card on any external site. -->
<div data-wrkhrs-brand="ideal-state"></div>
<script src="https://wrkhrs.co/embed/brand/ideal-state" async></script>

Call a prompt by id

curl -X POST https://wrkhrs.co/api/brand-spec/ideal-state/demo \
  -H 'content-type: application/json' \
  -d '{ "promptId": "hero_line", "variables": { "audience": "founders" } }'

Versioning

Spec version 1.0.0 as of 2017-09-21. Every release surfaces in the changelog.