Where the tool lives
Repository: github.com/CloudAxisAi/markdown-to-social
Install with npm install, then run node convert.js --file ./post.md. By default you get output/twitter-thread.txt, output/linkedin.txt, and output/instagram.txt. Use --stdout if you prefer piping instead of files.
What makes the output usable
Twitter / X: the script flattens your body copy, splits on word boundaries, and reserves space for 1/n numbering so each tweet stays under the limit — not a naive "split by H2 only" that breaks mid-paragraph.
LinkedIn: list lines become alternating 👉 / ✅ rows; links become label (url) so you can decide what to keep. Paragraphs are spaced for the "broetry" style that tends to read well in the feed.
Instagram: first ~150 words plus hashtag suggestions derived from tags: in frontmatter and section headings — still your job to trim for brand voice.
Honest limits
This is heuristic text shaping, not an LLM. It will not invent hooks, debate nuance, or match your company tone — it gives you paste-ready scaffolding. For fully automated multi-step publishing with agents, schedules, and integrations, that is a different product surface (see below).
When you want an agent, not a script
If your workflow is "research draft → edit in Notion → schedule to three channels → track replies," a one-shot CLI is not enough. CloudyBot is built for recurring execution with Specialists, file workspace, and optional cloud browser work — with hard caps so bills stay predictable. Try free → and see pricing for current task and browser-minute limits.
Further reading
- CloudyBot review: the AI agent that actually does work
- AI tools for freelancers (2026)
- Website change monitor CLI (2026)
Related reading
- How to use AI for SEO content without garbage
- How to automate social media posting with AI agents
- AI for content creators
Ready to automate this? CloudyBot can handle tasks like this on a schedule — with a real browser, memory, and WhatsApp delivery.
Try CloudyBot free →Free: 30 AI Tasks/month, no card required