How I Built an App With No CS Background (Firebase + React, AI-Assisted)

A few months ago, the author had no understanding of Cloud Functions or Firestore triggers. However, today, the anonymous accountability partner matching app Veyl is live in production, offering real-time chat, voice notes, payments, and matching logic. The author, having no traditional engineering background, built Veyl by collaborating closely with AI tools, gradually mastering Firebase and React to debug and enhance the app, tackling one issue at a time.

Veyl's core functionality revolves around users signing up with a pseudonym, being matched with someone pursuing a similar objective, and mutually holding each other accountable through text and voice check-ins. The anonymity is a crucial aspect of the app. The author faced challenging tasks, such as comprehending Firestore security rules, ensuring users can only interact with their own data and that sensitive fields (like premium status) are set server-side via Cloud Functions.

The author also dealt with atomic batch writes for the matching system to prevent double-matching, webhook signature verification using HMAC SHA512 for Paystack payments, and idempotency handling to prevent any double charging or re-triggering due to a webhook retry. Additionally, the author implemented voice note encoding in the browser (OfflineAudioContext → WAV) before upload to keep file sizes small without relying on server-side transcoding.

The author advises someone starting from scratch to acknowledge that AI tools are powerful in writing code but not as effective in teaching the underlying principles or troubleshooting when things go awry. The debugging process is where the actual learning occurs. Despite not skipping this crucial step, the author utilized AI tools to expedite their entry into the field. Veyl is accessible at veyl.space, and the author is available to discuss the stack, security decisions, or the AI-assisted workflow.