YT SEO Architect: The Architect Does the Work. You Get the Views.

This project started with a simple frustration: YouTube SEO tools give you scores and suggestions, but you still have to do all the work. Audit 50 videos? Manually rewrite 50 titles? No thanks. YT SEO Architect is different. It's built around Phronesis — an autonomous agent that scans your channel, finds underperforming videos, generates specific improvements, and can auto-apply the high-confidence ones. Hands off. While you sleep. Here's what one scan cycle does: 🔍 Tier 1 — Scan: Pulls your channel, scores every video across title, description, tags, CTR, and retention. No manual input. 💡 Tier 2 — Propose: All 8 propose agents fire unconditionally — title fixes, tag optimization, keyword opportunities, chapter suggestions, evergreen content, pre-upload checks, thumbnail analysis, A/B test variants. 38+ proposals per scan. ⚡ Tier 3 — Act: Tags at 90%+ confidence auto-apply to YouTube. Title/description changes queue to your inbox with side-by-side CURRENT → PROPOSED comparisons. Approve, edit, or skip. 🧪 A/B Title Tester: Start a test, it applies Variant A for 48h, auto-advances to Variant B, declare s a winner. All automated. 📈 12 programmatic SEO blog posts (1500+ words each) targeting YouTube creator keywords, auto-published with IndexNow submission to Bing/Yandex. Stack: Node.js 22 · Express 5 · Drizzle ORM + Neon Postgres · Groq/Gemini AI failover · YouTube Data API v3 · Vercel serverless The inbox is where the magic happens — every proposal shows you exactly what's changing and why, with confidence scores and expected value calculations. No black box. Would love feedback from creators, SEO folks, and anyone who's wrestled with YouTube's algorithm. What's your biggest YouTube SEO pain point? Drop it below and I'll tell you if Phronesis handles it. 🔗 yt-seo-architect.vercel.app