Build a Website for Browsers and AI Agents from a Shared Content Model
Description and use
A unified content model serves browsers (HTML) and AI agents (Markdown, APIs) with automatic revalidation, secure forms, and discoverable endpoints.
Problem
Need a single content source that serves both human visitors (HTML) and automated clients (Markdown, APIs) while keeping editorial flexibility and security.
Outcome
The platform delivers consistent, performant HTML to browsers and Markdown to AI agents from the same records, automates revalidation on content change, secures form input, and provides discoverable APIs via well‑known endpoints.
Workflow
- 1. Design 19 Slice Machine components in Prismic for page sections.
- 2. Store global content (navigation, SEO, announcements) as shared records.
- 3. Configure Next.js App Router for static generation, caching, and Draft Mode preview.
- 4. Sign form field definitions with HMAC and validate on server; restrict endpoints and add honeypot/timeouts.
- 5. Integrate Rippling for career listings, merging data in Next.js.
- 6. Enable HTTP content negotiation to serve Markdown via .md URLs.
- 7. Publish 6 discovery surfaces (.well-known/api-catalog, .well-known/mcp/server-card.json, .well-known/agent-skills/, .well-known/oauth-authorization-server, sitemap.md, /api/news/articles).
Evidence
Walden Robotics website uses 19 reusable content components and publishes 6 discovery surfaces for automated clients.
The platform connects Prismic content changes to revalidation, so editors work in the CMS while the application updates its generated output.
The platform signs every form field definition with HMAC before the form reaches the browser and validates the submitted data on the server.
- Record date: 2026-09-06T18:52:00.775Z