In-house Product

AI LinkedIn Content Engine — Auto-Posting Pipeline

An end-to-end AI system that generates, designs and schedules a full day of LinkedIn content automatically — pulling trending topics, writing the posts, building carousel PDFs and infographics, and queueing everything to publish, with almost no human input. Now live and open-sourced on GitHub.

Category
AI
Status
Live
Our role
AI systems design and automation engineering
Timeline
3 weeks
AI LinkedIn Content Engine — concept showing a scheduled LinkedIn post with an AI-built carousel and a fetch-to-schedule automation pipeline
AI LinkedIn Content Engine — concept showing a scheduled LinkedIn post with an AI-built carousel and a fetch-to-schedule automation pipeline

The problem

Staying consistently active on LinkedIn is brutal: it means finding fresh, relevant topics every day, writing posts that actually land, designing carousels and infographics, and then scheduling all of it on time. Done by hand it eats hours daily, and consistency collapses the moment teams get busy — which is exactly when visibility matters most.

What we built

Kalpabriksha built a multi-stage content pipeline (orchestrated with AI coding tools including the Antigravity IDE) that runs the whole workflow hands-free. Phase 1 fetches source material — trending Reddit threads, AI-news RSS feeds and a fresh data set for infographics. Phase 2 uses LLMs (via OpenRouter/Anthropic Claude) to generate the day's posts across several proven archetypes, governed by a written content doctrine and voice profile so everything stays on-brand. Phase 3 turns content into visual assets — 7-slide carousels rendered to PDF via Puppeteer, and HTML infographics screenshotted to PNG. Phase 4 delivers every asset to Slack for review, and Phase 5 drives browser automation to schedule the posts on LinkedIn across the coming days. The result is a repeatable, mostly autonomous "content factory".

Key capabilities

  • Automated topic sourcing from Reddit, AI-news RSS and live data sets
  • AI-written posts across multiple high-performing archetypes
  • On-brand output enforced by a content doctrine and voice profile
  • Auto-generated 7-slide carousels rendered to PDF (Puppeteer)
  • Auto-generated infographics (HTML → PNG screenshot)
  • Performance engine that models posts on real LinkedIn analytics
  • Slack delivery of all post text and visual assets for quick review
  • Browser-automated LinkedIn scheduling across multiple days
  • Configurable daily volume and posting cadence
  • Built and iterated with AI coding assistants (Antigravity IDE)

The outcome

Now live and open-sourced on GitHub. The full pipeline — topic fetching, AI generation, carousel/infographic rendering, Slack delivery and browser-automated LinkedIn scheduling — runs end-to-end for hands-free daily posting.