We built Zanderio by “vibe coding.” Here’s what worked (and what didn’t)

Hey Hunters — it’s Sheraz here, CTO & Co-Founder of Zanderio. We’ve been building by vibe coding: follow the energy, ship tiny, learn fast.

How we keep vibe ≠ chaos

• 48-hour cap per idea → demo or delete

• One metric per sprint (e.g., product discovery starts)

• “No meetings, only demos”

• Kill-switch + logs for every feature (easy rollback)

• Default: one helpful nudge/session; quiet after a “no”

What worked

• Momentum > motivation (daily win log)

• Intent-based nudges beat timers

• Auto-ingest of catalog/price/stock → setup < 10 min

What didn’t

• Over-generalizing early (we scrapped two “platform” attempts)

• Writing rules before hearing real shopper phrasing

• Stack (quick)

• React/Tailwind, Node, retrieval-augmented AI with guardrails, and Shopify/Woo hooks

Questions for you

• How do you keep vibe coding from drifting into bloat?

• Which constraint do you love (timebox, single metric, user count)?

• What’s your “delete it” signal?

I’ll reply to everyone plz do share your experience.