BurnLens — Free AI Spend Auditor
A no-login AI spend audit tool that tells startup founders and engineering teams exactly where their AI budget is leaking — with a BurnLens Score, savings breakdown, and a shareable PDF report in under 60 seconds.
View live✦ The Brief
Where it started
Build a free, zero-friction AI spend audit tool for startup founders and engineering managers. The goal: give any team an instant, honest picture of where their AI budget is leaking — no login, no sales call, no consultant needed. Just enter your tools, get your audit, share the link.
✦ The Thinking
Decisions before code
Three decisions shaped everything. First: hardcoded audit rules over AI-generated recommendations. The reasoning had to be deterministic and verifiable — a CFO needs to agree with the logic. Pure TypeScript functions are testable, auditable, and instant. Second: Next.js App Router over Pages Router. The shareable audit page needed server-side data fetching before render — critical for the Open Graph previews that make sharing actually work. Third: best-effort database writes. The audit runs entirely in memory. A failed Supabase write should never block a user from seeing their results. Resilience over perfect data capture.
✦ Building it
From sketch to ship
Next.js, TypeScript, Tailwind CSS, Supabase, Resend, and Vercel. The audit engine is pure TypeScript — deterministic functions that take a list of tools and return a verdict per tool, a savings calculation, and a 0–100 BurnLens Score. Supabase persists shareable audits with best-effort writes so the in-memory result is never blocked. The shareable URL is server-rendered for clean Open Graph previews when founders pass it around.
Stack
Screenshots
✦ What I'm proud of
The detail that mattered
The BurnLens Score — a single 0–100 efficiency rating that gives founders an instant gut-check without having to read a full report. The shareable public URL means audit results travel freely: a founder can send their numbers to their co-founder or CFO without any login friction on either end. Lighthouse scores: Desktop 97 / 96 / 100 / 100.
✦ What I'd do differently
The honest reflection
Add per-tool historical trend tracking so teams can watch whether their AI spend efficiency is improving or degrading month over month — not just a one-time snapshot.