All case studies
SaaS · audit engineWeb DevelopmentAI Automation

RoDevly

rodevly.com
Next.jsBetter AuthStripeAnthropic APIPostgreSQLLua plugin
RoDevly, SaaS · audit engine
Scroll to play

Brief

Roblox creators tune monetization by guesswork: prices copied from bigger games, game passes nobody can reach, shop screens that quietly do not exist. RoDevly is a linter for a game's economy. A Studio plugin reads the game's structure, its game passes, developer products and shop UI, and returns a 0 to 100 score with concrete findings in seconds. No account access, no scripts read, no player data required.

Decisions

01Rules that cannot be verified are left out, not faked

The engine ships 13 structural rules, each with a severity, a penalty and a concrete recommendation, and the score is plain arithmetic: 100 minus penalties. Anything that would require guessing about player behaviour stays out of the rulebook entirely. A smaller audit that is always defensible beats a bigger one that is sometimes wrong.

02Static analysis only, by design

The plugin reads structure: game passes, developer products, shop screens. It never reads script source and never touches an account. That boundary is what lets anyone run a public teardown of any game straight from the landing page, and it is why an audit finishes in seconds instead of waiting on gameplay data.

03Self-hosted auth instead of a branded widget

Auth runs on Better Auth inside the app itself: no external auth service, no third-party branding on the sign-in form, and the same library's Stripe plugin drives the subscriptions. One dependency fewer to pay for, and the login page belongs to the product.

04Quotas enforced where they cost money

The free and paid boundary is resolved server-side, straight from the subscription table, and the monthly AI report quota is counted inside the audit endpoint itself. Once a plan's cap is reached the Claude call is simply never made, so a runaway client cannot generate runaway API bills. The rule engine itself is never metered.

Result

Structural rules in the engine13
Performance (desktop)95
Largest Contentful Paint (desktop)1.4 s
Cumulative Layout Shift0
Total Blocking Time (desktop)0 ms
Best Practices100
SEO100
Performance (mobile)73

Mobile

Performance73
Accessibility96
Best Practices100
SEO100

Desktop

Performance95
Accessibility96
Best Practices100
SEO100

The 13 rules figure comes from the product's public rulebook. Mobile performance is 73 with a 7.3 s LCP: the landing runs its animated audit demo above the fold, and on a throttled CPU that hero is most of the cost. Deferring it is the next planned pass.

All figures are real Lighthouse measurements (v12.8.2) taken against the live site on July 2026. Nothing is estimated.

Screens · click to enlarge

Want results like these?

Tell me what you're building and I'll tell you honestly what it needs.

Start a project