Tags: saas, automation, indiehackers, productivity | I have a chronic illness that sometimes puts me on the couch for a full day

Tags: saas, automation, indiehackers, productivity | I have a chronic illness that sometimes puts me on the couch for a full day without warning. I also run a senior enablement role at a tech company, maintain six web apps, and have 17 published books. The math does not work unless some of the output runs without me. So in February, I spent two weeks building an autonomous content publishing system. The Problem: I was creating content for nine platforms every week: LinkedIn, Pinterest, Mastodon, Substack, Medium, Dev.to, IndieHackers, Quora, and my personal blog at chadtdyar.com. Each platform has different formatting, character limits, and audience expectations. I was spending 4-5 hours per week just reformatting the same ideas for different channels. The Architecture: Database is Supabase Postgres. A marketing_queue table stores every piece of content with fields for platform, day of week, body text, status, and source. Validation functions run on insert to catch formatting issues before they reach the publish queue. Content generation uses weekly batch sessions producing all nine platforms worth of content from a single story seed. I use Claude to help draft and adapt, but every story comes from real experience. Publishing pipeline: a scheduled task checks the queue daily and routes content to each platform. Short-form auto-approves. Long-form stays in review until I manually approve. Monitoring: PostHog tracks pageviews across all six apps and the blog. Sentry catches errors. What Worked: The Supabase validation layer is the unsung hero. Batching content from a single seed story keeps the week coherent across platforms. What I Would Do Differently: Start with fewer platforms. Build the review queue first. Invest in content validation earlier. The Result: The system published to 40 accounts on a day I could barely hold my phone. I am now building ContentForge at contentforgehq.com as a product version of this workflow. If you are building something similar, the key insight is: validate everything before it enters the queue, and never trust auto-publish without a safety layer.