AI-powered social media content platform that automates the entire content lifecycle — from viral topic research and structured post generation with customizable tone and intention, to AI image creation and scheduled auto-posting on LinkedIn and Instagram. Used personally to drive LinkedIn growth: 7,349 impressions, 953 members reached, 20+ new connections, and one reconnection converted into a paying customer.
Solopreneurs and small business owners face content burnout — they spend all day making high-stakes business decisions and have nothing left for social media. The result is either wasting time doing it themselves (poorly and inconsistently) or paying someone else (expensive and often generic). No existing tool owned the full content lifecycle: competitors generate text and stop, offer no visual content generation, and produce generic output with no strategic intent. There was no middle ground between expensive human help and inadequate DIY.
7,349 (+181% YoY)
Total impressions
5,168 (+1,267%)
90-day impressions
953 total (+514%)
Members reached
1 paying customer
Revenue impact

Built an end-to-end platform covering the entire content lifecycle. Designed three distinct user paths: "I know what to post" (direct generation), "I don't know what to post" (AI-researched trending topics from Reddit and social platforms), and "batch creation" (one focused session for a week of content). Implemented a 5-part structured post format (Hook, Context, Value, CTA, Hashtags), customizable tone profiles for brand consistency, AI image generation with selectable visual styles, and scheduled auto-posting via connected social accounts. Coordinated separate text and image AI pipelines through n8n webhooks, with Supabase edge functions handling the backend orchestration.
Development
Other
Databases
Automation
IDE
Competitor analysis of AI content tools (Jasper, Copy.ai, etc.), founder interviews with 2-3 solopreneurs about social media habits, and personal experience as the target user to identify gaps in the market
Designed three user paths (direct generation, idea discovery, batch creation), 5-part post structure, tone profile system, and multi-platform support. Built responsive layouts for desktop (sidebar + command palette) and mobile (bottom tabs, 44px+ touch targets)
Built the Post Generator with intention and tone controls, Ideas tab powered by AI agents scanning Reddit for viral topics, AI image generation integrated into the post flow, scheduling system with smart posting time recommendations, and auto-posting via OAuth-connected social accounts
Dogfooded the product as sole content creation tool for LinkedIn over several months (April–July 2025). Real posts to a real audience validated what features mattered most — batch scheduling and simple generation were the core, while advanced features like multi-variant generation were nice-to-have
“Ship incrementally, not all at once — the product became heavy with features, and some were never used. The simplest features (generate a post, schedule it, done) delivered the most value. Dogfooding is the best validation: daily personal use revealed what mattered faster and more honestly than any interview. Solve one problem well before solving five problems okay. This project taught the discipline of incremental building and validated that a solo-built AI agent can maintain a professional social media presence indistinguishable from manually crafted content.”