ContentForge

ContentForge is a standalone content automation engine I built to solve a real problem: manually creating social media posts for a spiritual brand across 4 platforms is not sustainable. The system reads spiritual content data (forms, stotras, stories, articles) from a TypeScript source, transforms it through platform-specific extractors, renders HTML templates into production-ready images via Puppeteer, and publishes on schedule.

The engine generates 8 distinct Instagram card types across 4 visual themes (classic, cosmic, puja-scene, shakti). Namavali cards display 4 divine names per card from 4 different namavalis totalling 302 names. Sahasranama cards pull from a 1,000-name dataset. Story carousels produce 7-slide posts: cover, context, action, climax, morals, insights, and CTA. Stotra verse cards and mantra cards round out the Instagram content.

Pinterest gets its own extraction layer: wisdom quotes from articles, form spotlight pins with deity images, moral teachings from stories, and ritual tip cards. All at 1000x1500 resolution with SEO-rich descriptions.

Twitter content follows different rules entirely. Tweets stay under 220 characters (leaving room for hashtags). The extractor scores story hooks by dramatic impact and picks the strongest opening lines. Article threads produce 7-8 tweet chains with hooks, key insights, and CTAs. Links go in auto-reply tweets only because embedding URLs in the main tweet body kills reach.

The scheduler manages a queue with statuses (DRAFT, QUEUED, PUBLISHING, PUBLISHED, FAILED) and respects safe daily limits: 2 posts/day on Instagram, 9 on Pinterest, 4 on Twitter. A cron job checks every 2 minutes for posts due to publish. Auto-fill scans output directories and queues unposted content with time slots.

The dashboard is a React 19 app with real-time SSE streaming for generation progress, queue management, OAuth token management, and content preview galleries. Platform integrations use Instagram Graph API v21.0, Pinterest API v5, and Twitter API v2.

I also built an AI-powered DM engine for Instagram. When users message the brand, Groq's LLM generates contextual replies based on a knowledge base of FAQs, form recommendations, and practice guidance.