Feature deep-dive

Headless API

Last updated

The Mentionwell Reader API is a clean, public, bearer-token-scoped REST endpoint. Pair it with the mentionwell npm package and you have typed loaders for Next.js (App Router + Pages), Astro, Remix, SvelteKit, Nuxt, React + Vite, and plain HTML in one install.

Endpoints

  • GET /v1/articles — paginated list with cursor.
  • GET /v1/articles/{slug} — full article including HTML, Markdown, JSON-LD.
  • GET /v1/articles/{slug}.md — markdown-only mirror for AI ingestion.
  • GET /v1/sites/{slug}/feed.xml — RSS.
  • GET /v1/sites/{slug}/feed.json — JSON Feed.
  • GET /v1/sites/{slug}/sitemap.xml — sitemap.
  • GET /v1/sites/{slug}/llms.txt — site-wide AI crawler manifest.

Authentication

Per-site reader keys (rotatable). Personal access tokens (mw_pat_*) for headless surfaces. HMAC-SHA256 signing on outbound webhooks.

Why headless

  • Your frontend keeps full control over rendering, caching, and routing.
  • No vendor lock-in on the frontend — switch frameworks any time.
  • ISR / on-demand revalidation works cleanly via webhook.
  • Works on the edge — Vercel, Cloudflare, Netlify functions all consume the API natively.

See also

All features Integrations API reference SDK reference CLI
66 beta testing now