What I learned fixing SEO and auth issues in a multilingual Next.js SaaS

I’ve been building Exarbi, a multilingual crypto arbitrage monitoring platform, and recently spent a lot of time fixing issues that looked small in the UI but mattered for crawlers and performance.

A few examples:

• Protected routes were redirecting users correctly, but crawlers could still see a 200 response before the client-side redirect.
• Noindex auth pages were inheriting canonical and hreflang metadata from public layouts.
• Public pages were linking directly to a protected dashboard route.
• A large client boundary in the public layout was making mostly static pages hydrate more JavaScript than necessary.
• Supporting eight languages made every metadata and routing decision more complicated.

The main lesson was that technical SEO in a modern SaaS is often an architecture problem, not a copywriting problem.

We moved authoritative auth checks to the server, separated public and protected metadata behavior, cleaned up locale-aware redirects, and started splitting large client components into smaller islands.

The product itself does not require exchange API keys, does not execute trades, and does not hold user funds. It is meant to help users inspect price differences together with transfer and risk conditions.

I’m still working through the positioning and onboarding.

For founders building technical B2B tools: what helped you explain a complex product without overwhelming first-time visitors?