Publish your AI blog to your own website.

FOUNDERREPLY_WORKSPACE_ID=<the uuid from your dashboard url>

# every published post, newest first

{
  "ok": true,
  "total": 24, "limit": 50, "offset": 0,
  "posts": [
    { "id", "slug", "title", "type",
      "keywords", "meta_description",
      "cover_image_url", "created_at",
      "published_at", "updated_at",
      "intent", "visibility" }, …
  ]
}

# one full article by slug

{
  "ok": true,
  "post": {
    …all list fields,
    "body_markdown": "…",
    "json_ld": [ Article, FAQPage? ]
  },
  "related": [
    { "title", "slug", "url"? }, …
  ]
}

Twelve fields per list item, and updated_atis the one to keep — it is what the sitemap’s lastmod is built from. related is scored on keyword overlap and gains an absolute url once you pass base_url — render those links server-side, because crawlers do not run your JavaScript.

Errors are never edge-cached, so a retry after a fix is answered fresh. Successful reads are, which is why a build that pulls the list once and then walks it usually stays well inside the rate limit.

Time-based (ISR)

Set a revalidate window on the consuming page — the site re-pulls in the background on the next request after the window lapses. e.g. revalidate: 300 = at most 5 min stale.Simplest. Good default.

Scheduled rebuild

Run a cron (GitHub Action / Cloudflare cron / Vercel cron) that triggers a rebuild on a fixed cadence — nightly, hourly, whatever you choose. The site is 100% static between builds.Predictable, fully static.

On-publish trigger

Hit your host’s deploy hook the moment you mark a post published, so the new article goes live within a build cycle instead of waiting for a timer.Freshest. Event-driven.

// re-pull at most every 5 minutes
export const revalidate = 300

const WS = process.env.FOUNDERREPLY_WORKSPACE_ID
const BASE = "https://founderreply.com/api/public/blog"

// one route per published slug
export async function generateStaticParams() {
  const { posts } = await fetch(`${BASE}?workspace_id=${WS}`).then(r => r.json())
  return posts.map((p) => ({ slug: p.slug }))
}

// render one article
export default async function Page({ params }) {
  const { slug } = await params
  const { post } = await fetch(`${BASE}/${slug}?workspace_id=${WS}`).then(r => r.json())
  return (
    <article>
      <script
        type="application/ld+json"
        dangerouslySetInnerHTML={{ __html: JSON.stringify(post.json_ld) }}
      />
      <Markdown>{post.body_markdown}</Markdown>
    </article>
  )
}

Use any markdown renderer (react-markdown, marked, MDX). The same pattern works in SvelteKit, Nuxt, Astro, or a plain build script — it is two GET requests. Pass base_url on the detail call and the JSON-LD comes back carrying your canonical URL, so you do not have to patch it in.

# the two XML surfaces — rooted at your blog base
curl "https://founderreply.com/api/public/blog/sitemap.xml?workspace_id=$WS&base_url=$BLOG_BASE"
curl "https://founderreply.com/api/public/blog/rss.xml?workspace_id=$WS&base_url=$BLOG_BASE"

Leave base_url off and both fall back to the blog base configured on your primary SEO site. With neither, they return 400 — better than guessing at a domain and emitting a sitemap full of URLs that do not resolve.