Skip to content

The Postgres development platform. Supabase gives you a dedicated Postgres database to build your web, mobile, and AI applications.

feat(www): serve blog, customers, events as .md for AI agents (#45403)

Pamela Chiabaabcb189c5f
diff --git a/apps/www/app/blog/[slug]/page.tsx b/apps/www/app/blog/[slug]/page.tsx
index e2f68cd513b..6e6fc8fb88c 100644
--- a/apps/www/app/blog/[slug]/page.tsx
+++ b/apps/www/app/blog/[slug]/page.tsx
@@ -2,8 +2,8 @@ import type { Metadata } from 'next'
import { draftMode } from 'next/headers'
import BlogPostClient from './BlogPostClient'
-import { SITE_ORIGIN } from '@/lib/constants'
import { getAbsoluteBlogSocialImage } from '@/lib/blog-images'
+import { SITE_ORIGIN } from '@/lib/constants'
import { getAllPostSlugs, getPostdata, getSortedPosts } from '@/lib/posts'
import type { Blog, BlogData, PostReturnType } from '@/types/post'
@@ -53,6 +53,11 @@ export async function generateMetadata({ params }: { params: Promise<Params> }):
return {
title: blogPost.title,
description: blogPost.description,
+ alternates: {
+ types: {
+ 'text/markdown': `/blog/${slug}.md`,
+ },
+ },
openGraph: {
title: blogPost.title,
description: blogPost.description,
diff --git a/apps/www/package.json b/apps/www/package.json
index 219f472836a..30503d95897 100644
--- a/apps/www/package.json
+++ b/apps/www/package.json
@@ -111,8 +111,10 @@
"api-types": "workspace:*",
"postcss": "catalog:",
"react-hook-form": "^7.71.2",
+ "remark-mdx": "^3.0.1",
"tailwindcss": "catalog:",
"tsconfig": "workspace:*",
+ "unist-util-visit": "^5.1.0",
"unist-util-visit-parents": "5.1.3",
"uuid": "^9.0.1",
"vite-tsconfig-paths": "catalog:",
diff --git a/apps/www/scripts/generateMdContent.mjs b/apps/www/scripts/generateMdContent.mjs
index 053408a6d26..378b18cefbf 100644
--- a/apps/www/scripts/generateMdContent.mjs
+++ b/apps/www/scripts/generateMdContent.mjs
@@ -1,31 +1,91 @@
// @ts-check
/**
- * Scans content/md/ and emits a TypeScript module exporting the markdown
- * content as a Map plus the slug allowlist as a Set. Static imports of the
- * generated file make the content traceable by @vercel/nft so it ends up in
- * the serverless bundle without runtime fs.readFile.
- *
- * Drop a new .md file in content/md/ and the route handler picks it up on
- * the next content:build — no constant to edit.
+ * Scans content/md/ + _blog/ + _customers/ + _events/ and emits a TypeScript
+ * module exporting MD_CONTENT (slug → markdown) and MD_PAGES (allowlist Set).
+ * The static import keeps content traceable by @vercel/nft, no runtime fs reads.
*/
import { promises as fs } from 'fs'
import path from 'path'
import { fileURLToPath } from 'url'
+import matter from 'gray-matter'
+
+import { mdxBodyToMarkdown } from './lib/mdxToMarkdown.mjs'
const __dirname = path.dirname(fileURLToPath(import.meta.url))
-const contentDir = path.join(__dirname, '../content/md')
-const outputPath = path.join(__dirname, '../app/api-v2/md/content.generated.ts')
+const wwwDir = path.join(__dirname, '..')
+const contentDir = path.join(wwwDir, 'content/md')
+const outputPath = path.join(wwwDir, 'app/api-v2/md/content.generated.ts')
+
+// Matches lib/posts.tsx FILENAME_SUBSTRING — strips YYYY-MM-DD- (11 chars).
+const DATE_PREFIX = 11
-// 'pricing' is served dynamically via generatePricingContent() instead of
-// from a .md file; it still needs to be in MD_PAGES so middleware rewrites
-// /pricing.md to the API route.
+// Slugs handled by a dynamic generator in the route handler. Listed here so
+// MD_PAGES still includes them (middleware relies on the allowlist).
const DYNAMIC_SLUGS = ['pricing']
+function pickFields(data, fields) {
+ const picked = {}
+ for (const field of fields) {
+ const value = data[field]
+ if (value == null || value === '' || (Array.isArray(value) && value.length === 0)) continue
+ picked[field] = value
+ }
+ return picked
+}
+
+const MDX_SECTIONS = [
+ {
+ dir: '_blog',
+ urlPrefix: 'blog',
+ stripDatePrefix: true,
+ frontmatterFields: ['title', 'description', 'author', 'date', 'tags', 'categories'],
+ },
+ {
+ dir: '_customers',
+ urlPrefix: 'customers',
+ stripDatePrefix: false,
+ frontmatterFields: [
+ 'name',
+ 'title',
+ 'description',
+ 'company_url',
+ 'industry',
+ 'region',
+ 'company_size',
+ 'supabase_products',
+ 'date',
+ ],
+ },
+ {
+ dir: '_events',
+ urlPrefix: 'events',
+ stripDatePrefix: true,
+ frontmatterFields: [
+ 'title',
+ 'subtitle',
+ 'description',
+ 'type',
+ 'date',
+ 'end_date',
+ 'timezone',
+ 'duration',
+ 'onDemand',
+ ],
+ skipIf: (data) => data.disable_page_build === true,
+ },
+]
+
async function collectMdFiles(dir, prefix = '') {
const results = []
- const dirents = await fs.readdir(dir, { withFileTypes: true })
+ let dirents
+ try {
+ dirents = await fs.readdir(dir, { withFileTypes: true })
+ } catch (err) {
+ if (err.code === 'ENOENT') return results
+ throw err
+ }
for (const dirent of dirents) {
const slug = prefix ? `${prefix}/${dirent.name}` : dirent.name
if (dirent.isDirectory()) {
@@ -37,56 +97,117 @@ async function collectMdFiles(dir, prefix = '') {
return results
}
-// homepage first (it's the site overview), then everything else alphabetical.
+// _events double-underscore filenames (`2024-08-30__launch-...`) produce slugs
+// with a leading underscore. The HTML page is at `/events/_launch-...`, so the
+// .md slug must keep the underscore to match.
+function deriveSlug(filename, stripDatePrefix) {
+ const base = filename.replace(/\.mdx$/, '')
+ return stripDatePrefix ? base.substring(DATE_PREFIX) : base
+}
+
+async function ingestMdxSection(section) {
+ const dir = path.join(wwwDir, section.dir)
+ let filenames
+ try {
+ filenames = await fs.readdir(dir)
+ } catch (err) {
+ if (err.code === 'ENOENT') {
+ console.warn(` ⚠ Section directory missing: ${section.dir}`)
+ return []
+ }
+ throw err
+ }
+
+ const mdxFilenames = filenames.filter((f) => f.endsWith('.mdx'))
+ const results = await Promise.all(
+ mdxFilenames.map(async (filename) => {
+ const fullPath = path.join(dir, filename)
+ try {
+ const raw = await fs.readFile(fullPath, 'utf-8')
+ const parsed = matter(raw)
+ const data = parsed.data ?? {}
+ if (section.skipIf?.(data)) return null
+
+ const slug = `${section.urlPrefix}/${deriveSlug(filename, section.stripDatePrefix)}`
+ const body = await mdxBodyToMarkdown(parsed.content)
+ const content = matter.stringify(body, pickFields(data, section.frontmatterFields))
+ return { slug, content }
+ } catch (err) {
+ throw new Error(`${section.dir}/${filename}: ${err.message}`, { cause: err })
+ }
+ })
+ )
+ return results.filter((entry) => entry !== null)
+}
+
function sortSlugs(a, b) {
if (a === 'homepage') return -1
if (b === 'homepage') return 1
return a.localeCompare(b)
}
-const slugs = (await collectMdFiles(contentDir)).sort(sortSlugs)
+const staticSlugs = (await collectMdFiles(contentDir)).sort(sortSlugs)
-if (slugs.length === 0) {
+if (staticSlugs.length === 0) {
console.error('❌ No .md files found in content/md/')
process.exit(1)
}
-// A static file with the same slug as a dynamic generator would land in both
-// MD_CONTENT and the dynamic append path, so /llms-full.txt would emit it
-// twice. Fail the build instead of shipping a duplicate.
-const collisions = slugs.filter((s) => DYNAMIC_SLUGS.includes(s))
-if (collisions.length > 0) {
+const staticEntries = await Promise.all(
+ staticSlugs.map(async (slug) => ({
+ slug,
+ content: await fs.readFile(path.join(contentDir, `${slug}.md`), 'utf-8'),
+ }))
+)
+
+const mdxEntries = []
+for (const section of MDX_SECTIONS) {
+ console.log(`📚 Ingesting ${section.dir}...`)
+ const sectionEntries = await ingestMdxSection(section)
+ console.log(` ${sectionEntries.length} entries`)
+ mdxEntries.push(...sectionEntries)
+}
+
+const allEntries = [...staticEntries, ...mdxEntries]
+
+const dynamicCollisions = staticSlugs.filter((s) => DYNAMIC_SLUGS.includes(s))
+if (dynamicCollisions.length > 0) {
console.error(
- `❌ Slug collision: [${collisions.join(', ')}] is reserved for a dynamic generator. ` +
- `Remove the corresponding file from content/md/ or update DYNAMIC_SLUGS.`
+ `❌ Slug collision: [${dynamicCollisions.join(', ')}] reserved for a dynamic generator.`
)
process.exit(1)
}
-const entries = []
-const errors = []
-for (const slug of slugs) {
- const filePath = path.join(contentDir, `${slug}.md`)
- try {
- const content = await fs.readFile(filePath, 'utf-8')
- entries.push(` [${JSON.stringify(slug)}, ${JSON.stringify(content)}]`)
- } catch (err) {
- errors.push(`${slug}.md: ${err.message}`)
- }
+const sectionRoots = MDX_SECTIONS.map((s) => s.urlPrefix)
+const sectionRootCollisions = [...staticSlugs, ...DYNAMIC_SLUGS].filter((s) =>
+ sectionRoots.includes(s)
+)
+if (sectionRootCollisions.length > 0) {
+ console.error(
+ `❌ Section-root collision: [${sectionRootCollisions.join(', ')}] shadows an MDX section.`
+ )
+ process.exit(1)
}
-if (errors.length > 0) {
- console.error('❌ Failed to read .md files:')
- errors.forEach((e) => console.error(` ${e}`))
- process.exit(1)
+const seen = new Set()
+for (const entry of allEntries) {
+ if (seen.has(entry.slug)) {
+ console.error(`❌ Duplicate slug emitted: ${entry.slug}`)
+ process.exit(1)
+ }
+ seen.add(entry.slug)
}
-const allSlugs = [...slugs, ...DYNAMIC_SLUGS]
-const pageEntries = allSlugs.map((s) => ` ${JSON.stringify(s)}`).join(',\n')
+const contentEntries = allEntries
+ .map((e) => ` [${JSON.stringify(e.slug)}, ${JSON.stringify(e.content)}]`)
+ .join(',\n')
+
+const allPageSlugs = [...allEntries.map((e) => e.slug), ...DYNAMIC_SLUGS]
+const pageEntries = allPageSlugs.map((s) => ` ${JSON.stringify(s)}`).join(',\n')
const output = `// AUTO-GENERATED by scripts/generateMdContent.mjs — do not edit
export const MD_CONTENT = new Map<string, string>([
-${entries.join(',\n')},
+${contentEntries},
])
export const MD_PAGES = new Set<string>([
@@ -95,4 +216,4 @@ ${pageEntries},
`
await fs.writeFile(outputPath, output, 'utf-8')
-console.log(`✅ Generated ${outputPath} (${entries.length} files, ${allSlugs.length} pages)`)
+console.log(`✅ Generated ${outputPath} (${allEntries.length} files, ${allPageSlugs.length} pages)`)
diff --git a/apps/www/scripts/lib/__tests__/mdxToMarkdown.test.mjs b/apps/www/scripts/lib/__tests__/mdxToMarkdown.test.mjs
new file mode 100644
index 00000000000..76ff2c551e5
--- /dev/null
+++ b/apps/www/scripts/lib/__tests__/mdxToMarkdown.test.mjs
@@ -0,0 +1,89 @@
+// @ts-check
+import { describe, expect, it } from 'vitest'
+
+import { mdxBodyToMarkdown } from '../mdxToMarkdown.mjs'
+
+describe('mdxBodyToMarkdown', () => {
+ it('preserves plain markdown unchanged', async () => {
+ const out = await mdxBodyToMarkdown('# Title\n\nA **bold** word.')
+ expect(out).toContain('# Title')
+ expect(out).toContain('**bold**')
+ })
+
+ it('preserves code fences containing JSX literally', async () => {
+ const src = [
+ '```tsx',
+ 'function App() {',
+ ' return <Quote caption="x">hi</Quote>',
+ '}',
+ '```',
+ ].join('\n')
+ const out = await mdxBodyToMarkdown(src)
+ expect(out).toContain('<Quote caption="x">hi</Quote>')
+ expect(out).toContain('```tsx')
+ })
+
+ it('rewrites <Img /> to a markdown image', async () => {
+ const out = await mdxBodyToMarkdown('<Img src="/x.png" alt="X" />')
+ expect(out.trim()).toBe('![X](/x.png)')
+ })
+
+ it('rewrites <Img /> without alt', async () => {
+ const out = await mdxBodyToMarkdown('<Img src="/x.png" />')
+ expect(out.trim()).toBe('![](/x.png)')
+ })
+
+ it('rewrites <Admonition> to a labeled blockquote', async () => {
+ const out = await mdxBodyToMarkdown('<Admonition type="note">\n\nHeads up.\n\n</Admonition>')
+ expect(out).toMatch(/^>\s+\*\*NOTE\*\*/m)
+ expect(out).toContain('Heads up.')
+ })
+
+ it('rewrites <Quote> with caption to a blockquote with em-dash attribution', async () => {
+ const out = await mdxBodyToMarkdown(
+ '<Quote caption="Jane Doe, CEO">\n\nIt was great.\n\n</Quote>'
+ )
+ expect(out).toContain('> It was great.')
+ expect(out).toContain('*— Jane Doe, CEO*')
+ })
+
+ it('rewrites <Link href> to a markdown link', async () => {
+ const out = await mdxBodyToMarkdown('<Link href="/x">click</Link>')
+ expect(out.trim()).toBe('[click](/x)')
+ })
+
+ it('rewrites <Subtitle> and <Badge> to bold', async () => {
+ const out = await mdxBodyToMarkdown('<Subtitle>foo</Subtitle> <Badge>bar</Badge>')
+ expect(out).toContain('**foo**')
+ expect(out).toContain('**bar**')
+ })
+
+ it('drops <Avatar> entirely', async () => {
+ const out = await mdxBodyToMarkdown('<Avatar src="/x" /> rest')
+ expect(out).not.toContain('Avatar')
+ expect(out).toContain('rest')
+ })
+
+ it('flattens unknown inline JSX to text, preserving inner content', async () => {
+ const out = await mdxBodyToMarkdown('Some text with <Mystery>kept</Mystery> inline.')
+ expect(out).toContain('kept')
+ expect(out).toContain('Some text with')
+ expect(out).toContain('inline.')
+ expect(out).not.toContain('<Mystery')
+ expect(out).not.toContain('</Mystery')
+ })
+
+ it('strips MDX comments {/* ... */}', async () => {
+ const out = await mdxBodyToMarkdown('hello {/* internal note */} world')
+ expect(out).toContain('hello')
+ expect(out).toContain('world')
+ expect(out).not.toContain('internal')
+ })
+
+ it('strips orphan mdxjsEsm imports at the top level', async () => {
+ const src = "import X from 'y'\n\n# Title"
+ const out = await mdxBodyToMarkdown(src)
+ expect(out).toContain('# Title')
+ expect(out).not.toContain('import')
+ })
+})
diff --git a/apps/www/scripts/lib/mdxToMarkdown.mjs b/apps/www/scripts/lib/mdxToMarkdown.mjs
new file mode 100644
index 00000000000..c17e844be7c
--- /dev/null
+++ b/apps/www/scripts/lib/mdxToMarkdown.mjs
@@ -0,0 +1,114 @@
+// @ts-check
+
+import { remark } from 'remark'
+import remarkGfm from 'remark-gfm'
+import remarkMdx from 'remark-mdx'
+import { SKIP, visit } from 'unist-util-visit'
+
+function getJsxAttr(node, name) {
+ const attr = (node.attributes ?? []).find((a) => a.type === 'mdxJsxAttribute' && a.name === name)
+ return typeof attr?.value === 'string' ? attr.value : ''
+}
+
+function flattenChildText(nodes) {
+ const parts = []
+ for (const node of nodes) {
+ if (node.type === 'text' || node.type === 'inlineCode') parts.push(node.value)
+ else if (node.children) parts.push(flattenChildText(node.children))
+ }
+ return parts.join('')
+}
+
+const text = (value) => ({ type: 'text', value })
+const para = (children) => ({ type: 'paragraph', children })
+const strong = (children) => ({ type: 'strong', children })
+const emph = (children) => ({ type: 'emphasis', children })
+const blockquote = (children) => ({ type: 'blockquote', children })
+
+// Unknown JSX must not become a `paragraph` (block) inside an inline parent.
+function unwrap(node) {
+ const children = node.children ?? []
+ if (node.type === 'mdxJsxTextElement') return text(flattenChildText(children))
+ return children.length > 0 ? para(children) : text('')
+}
+
+const REWRITES = {
+ Img: (node) => {
+ const src = getJsxAttr(node, 'src')
+ return src ? { type: 'image', url: src, alt: getJsxAttr(node, 'alt'), title: null } : text('')
+ },
+ img: (node) => REWRITES.Img(node),
+
+ Admonition: (node) => {
+ const label = (getJsxAttr(node, 'type') || 'note').toUpperCase()
+ return blockquote([para([strong([text(label)])]), ...(node.children ?? [])])
+ },
+
+ Quote: (node) => {
+ const caption = getJsxAttr(node, 'caption')
+ const attribution = caption ? [para([emph([text(`— ${caption}`)])])] : []
+ return blockquote([...(node.children ?? []), ...attribution])
+ },
+
+ BlogCollapsible: (node) => {
+ const title = getJsxAttr(node, 'title')
+ const heading = title ? [para([strong([text(title)])])] : []
+ return blockquote([...heading, ...(node.children ?? [])])
+ },
+
+ Link: (node) => {
+ const href = getJsxAttr(node, 'href')
+ if (!href) return unwrap(node)
+ return { type: 'link', url: href, title: null, children: node.children ?? [] }
+ },
+ a: (node) => REWRITES.Link(node),
+
+ Subtitle: (node) => strong(node.children ?? []),
+ Badge: (node) => strong(node.children ?? []),
+
+ Avatar: (node) => (node.type === 'mdxJsxTextElement' ? text('') : para([])),
+}
+
+function rewriteJsxNode(node) {
+ return REWRITES[node.name]?.(node) ?? unwrap(node)
+}
+
+function transformMdxNodes() {
+ return (tree, file) => {
+ const unknownNames = new Set()
+ visit(tree, (node, index, parent) => {
+ if (parent == null || index == null) return
+
+ if (
+ node.type === 'mdxFlowExpression' ||
+ node.type === 'mdxTextExpression' ||
+ node.type === 'mdxjsEsm'
+ ) {
+ parent.children.splice(index, 1)
+ return [SKIP, index]
+ }
+
+ if (node.type !== 'mdxJsxFlowElement' && node.type !== 'mdxJsxTextElement') return
+
+ if (node.name && !(node.name in REWRITES)) unknownNames.add(node.name)
+ parent.children[index] = rewriteJsxNode(node)
+ return [SKIP, index]
+ })
+ if (unknownNames.size > 0) {
+ file.message(`Unknown JSX components stripped: ${[...unknownNames].sort().join(', ')}`)
+ }
+ }
+}
+
+const processor = remark().use(remarkMdx).use(remarkGfm).use(transformMdxNodes).data('settings', {
+ bullet: '-',
+ fences: true,
+ incrementListMarker: false,
+ rule: '-',
+})
+
+export async function mdxBodyToMarkdown(rawBody) {
+ const file = await processor.process(rawBody)
+ for (const msg of file.messages) console.warn(` ⚠ ${msg.message}`)
+ return String(file)
+}
diff --git a/pnpm-lock.yaml b/pnpm-lock.yaml
index 10af73a9147..184a2b27ed6 100644
--- a/pnpm-lock.yaml
+++ b/pnpm-lock.yaml
@@ -1855,12 +1855,18 @@ importers:
react-hook-form:
specifier: ^7.71.2
version: 7.72.1([email protected])
+ remark-mdx:
+ specifier: ^3.0.1
+ version: 3.0.1([email protected])
tailwindcss:
specifier: 'catalog:'
version: 4.2.4
tsconfig:
specifier: workspace:*
version: link:../../packages/tsconfig
+ unist-util-visit:
+ specifier: ^5.1.0
+ version: 5.1.0
unist-util-visit-parents:
specifier: 5.1.3
version: 5.1.3
@@ -34099,7 +34105,7 @@ snapshots:
hast-util-heading-rank: 3.0.0
hast-util-is-element: 3.0.0
unified: 11.0.5
- unist-util-visit: 5.0.0
+ unist-util-visit: 5.1.0
@@ -34142,7 +34148,7 @@ snapshots:
github-slugger: 2.0.0
hast-util-heading-rank: 3.0.0
hast-util-to-string: 3.0.0
- unist-util-visit: 5.0.0
+ unist-util-visit: 5.1.0
dependencies: