← knowledge.oriz.in

Blog strategy 2026-07-01 — one source, multi-target cross-posting

decision blogcontentcross-postmdxmarkdown

Blog strategy 2026-07-01

Decision

Single source = CommonMark + YAML frontmatter (.md) in pages-blog-app. MDX reserved for the Astro build only — never the transport format. Cross-post via API to 4 write-capable platforms; canonical URL always blog.oriz.in.

Source format

---
title: "…"
description: "…"
date: 2026-07-01
tags: [tag1, tag2]
canonical_url: https://blog.oriz.in/posts/<slug>
cover_image: https://…
category: tech        # tech | personal | reviews
---

# Body in CommonMark GFM.

MDX (.mdx) allowed only in the Astro repo; the cross-post engine strips JSX before dispatch. MDsvex, MDXPortable rejected — not accepted by any target API.

Cross-post targets

Platform API Auth Free tier Format accepted
Dev.to POST /api/articles [1] api-key header Unlimited body_markdown + canonical_url
Hashnode GraphQL publishPost [2] PAT Unlimited contentMarkdown + originalArticleURL
Ghost (self-host) POST /admin/posts/?source=html [3] JWT Self-host HTML (Lexical internal)
WordPress POST /wp/v2/posts [4] App password Self-host / .com free HTML (MD → HTML on push)
Bluesky AT Proto app-password Free Plain text + link
Mastodon REST POST /api/v1/statuses Bearer Free Plain text
Medium ❌ Closed to new integrations Jan 2025 [5] — — —
Substack ❌ No public API [6] — — (unofficial libs only)
LinkedIn POST /rest/posts [7] OAuth2 Rate-limited Text + link (no article API for partners)

Medium + Substack: skip. LinkedIn: draft-only queue to chirag127/oriz-drafts GH Issues per blog-cross-post-strategy.

Category count

Three flat categories on one domain: tech / personal / reviews. Not per-app blogs. Not single-bucket. Reason: 2026 SEO consensus = topic-cluster on one domain beats micro-domains for authority [8][9]; three clusters = enough separation for feed filtering without diluting the pillar page.

Should there be a blog website?

Yes — blog.oriz.in stays canonical + public. Cross-posts point back via canonical_url. knowledge/ remains private/GitHub-only (agent memory, not blog). All three (site + cross-post + knowledge) coexist — different audiences.

Publishing workflow

  1. Author .md in pages-blog-app/src/content/blog/YYYY-MM-DD-slug.md.
  2. git push → Astro builds → blog.oriz.in/posts/<slug> live + RSS updated.
  3. GH Actions on tag v* calls @chirag127/omni-publish.
  4. Engine reads RSS <guid>, diffs against state.json, POSTs to Dev.to + Hashnode + Bluesky + Mastodon (+ Ghost/WP if configured) with canonical_url = blog.oriz.in/....
  5. LinkedIn/X/Reddit → draft issues in chirag127/oriz-drafts for manual review.
  6. state.json commit-back records adapter results; failures dead-letter with backoff.

Sources

  1. https://developers.forem.com/api/v1#tag/articles/operation/createArticle
  2. https://apidocs.hashnode.com (2.0 GraphQL, publishPost mutation)
  3. https://ghost.org/docs/admin-api/#creating-a-post
  4. https://developer.wordpress.org/rest-api/reference/posts/
  5. https://help.medium.com/hc/en-us/articles/213480228 — "will not be issuing any new integration tokens"
  6. https://github.com/mvanhorn/printing-press-library — "Substack has no public API"
  7. https://learn.microsoft.com/en-us/linkedin/marketing/community-management/shares/posts-api
  8. https://cicero.studio/en/blog/topic-cluster-seo (2026)
  9. https://vaza.ai/blog/mdx-vs-markdown-seo — "MDX and Markdown produce identical HTML"