HittinCorners: 869 Commits, a Full Stack Migration, and Then the Domain Just Stopped

HittinCorners was a "Yelp for DeFi" directory — 237 quality-reviewed protocols across 14 blockchain networks, platform comparisons, category and chain browsing, affiliate-tracked outbound links. It ran at hittincorners.com. As of this writing, that domain no longer resolves.

This is the biggest single bet in this whole portfolio, and it's the one with the least ambiguous ending, so it gets an honest postmortem instead of a launch post.

Crypto directory sites are a crowded category, and most of them are thin — a scraped list with an affiliate link bolted on. The idea was that a genuinely curated, quality-reviewed directory (237 platforms, hand-organized into 20 categories across 14 chains, real comparison pages and how-to guides) would out-compete the thin versions on trust and SEO both. That's a content-depth bet, and it got real investment behind it: 769 commits over more than five months on the original build alone.

What Got Built — Twice

The first version was a full production stack: Next.js 15, PayloadCMS as the content layer, Postgres with connection pooling, server-cached reads, tracked affiliate redirects through /go/[slug], a hardened multi-stage Docker deploy. Not a prototype — a real production application with a production database dump to show for it.

Then, five months in, it got rebuilt. Not patched — rebuilt, onto a completely different stack: Astro on Cloudflare Workers, D1 instead of Postgres, KV for sessions. The rewrite kept the same domain and the same core proposition (protocol reviews, market data, comparisons) but used the migration as the moment to add something the original never had: a homepage community feed — X-login or magic-link auth, post types for Discussion/Link/Ask/Show, voting, comment threads tied back into every platform and article page. The pitch for the rebuild was explicit in its own docs: editorial content and user-generated content should reinforce each other, reviews build trust, the feed builds proof and retention.

The Struggle

Migrating a live product's entire backend while simultaneously shipping a net-new feature is a genuinely hard sequencing problem — normal advice says do one or the other, not both at once. This did both at once: Postgres to D1, Next.js to Astro, and a full community feed with spam detection, sort algorithms, and mobile-specific UX (sticky reply bars, pull-to-refresh) landing in the same 100-commit push. A hundred commits in under a month for that scope is fast, and fast migrations of live products are exactly where things quietly break.

The Logic Behind the Feed

Directory content earns trust and search visibility slowly. A community feed was supposed to make that payoff compound faster — user posts and discussion threads feeding back into SEO and retention instead of the site depending purely on editorial output. The redesign collapsed multiple entry points (/explore, /feed, /talk) into one front door for exactly that reason: fewer places for intent to leak out before it converts.

Where It Stands

Offline. hittincorners.com doesn't resolve anymore. What's genuinely worth taking from this one isn't a growth number — it's that a serious content and infrastructure investment (869 commits combined, a real production database, a full stack migration executed live) doesn't guarantee survival, and that's worth saying plainly rather than quietly deleting the project from the list. Some bets don't make it. This was one of the bigger ones, and it still didn't.

Tried to pull its historical Search Console numbers for this update — permission denied. Can't even get clean access to the corpse's search data anymore, which feels like the appropriate final word on this one.