Skip to main content

Tech Debt Check: code audit for JavaScript and TypeScript apps, AI-generated or legacy

Give me your tech debt. I'll tell you where you're at.

I take a brief look at your codebase and tell you, honestly, how healthy it is. Free, in the order requests come in. Can't wait? Fast-track it and get the report within 48 hours.

Request a Tech Debt Check →See pricing

Łukasz Adamczyk · 14 years building and rescuing production systems. I maintain qoq, an open-source code quality toolkit.

Brief look · samplefictional project
58/ 100
FairWorks on prod, but it's getting expensive to change.
Security41
Dependencies46
Tests63
Consistency55
Maintainability72
CriticalSession token readable from JavaScript
Quick win · XSPatch the 3 vulnerable packages.
See the full sample report →
First come, first served

Legacy code audit and AI-generated code review

Legacy systems collect debt over years. AI-assisted code can collect it in weeks: missing tests, outdated or unnecessary dependencies, inconsistent patterns, security gaps, often faster than a team notices. The check looks at the code as it is, regardless of who wrote it.

AI-generated code

AI-built apps: will it survive real users?

AI-generated tech debt is the cost of code that works in a demo but was never designed: no tests, packages nobody chose on purpose, three patterns for the same job, and security gaps the model did not think about. It is not hypothetical. In Veracode's 2025 study, AI models introduced an OWASP Top 10 vulnerability in 45% of coding tasks. A Stanford study found that developers with an AI assistant wrote less secure code and were more confident it was secure.

If you built with Cursor, Claude Code, Copilot, Lovable or Bolt, the check tells you, in plain words, what will break first when real users arrive.

More on the vibe code audit →
Legacy code

Legacy JavaScript and TypeScript: what will the next upgrade cost?

Legacy debt builds up slowly: a framework two majors behind, a Node version past end of life, a test suite nobody trusts, and one module only one person understands. Each one makes features slower and upgrades scarier.

The check shows where the cost sits, what to fix before the next upgrade, and what can wait, in a form you can take to your manager or your board. I have spent 14 years on React, Node and the systems around them, so the advice comes from migrations I have actually led.

What the free technical debt assessment covers

Quality score
One number for where the codebase stands today, and why.
Headline risks
What can hurt you now or as the code grows, starting with security.
Quick wins
Changes with the best payoff for the least effort, whatever the code's age.

See what you get: sample brief look report · sample full audit report

What does the check inspect?

Five areas, each scored 0–100. Scores combine automated checks (qoq, npm audit, type and test coverage) with my reading of the code.

Security
Authentication and access checks, exposed secrets, input validation, known vulnerabilities in dependencies.
Dependencies
Outdated, deprecated, unused and duplicate packages, and how many major versions behind you are.
Tests
What is covered, what is not, and whether the risky paths (login, payments, data export) are among the gaps.
Consistency
Competing ways of doing the same job: data fetching, state, styling, error handling.
Maintainability
Types, module boundaries, and the hot-spot files everyone is afraid to touch.
0–39 at risk40–59 fair60–79 healthy80+ strong

The full audit adds delivery and infrastructure: CI, deploys, hosting, secrets handling and observability, read from config and docs.

How it works and what it costs

Pick what fits. Everything is covered by an NDA before any code changes hands.

Process

  1. 01RequestTwo minutes. Your package.json if you have one, and a few lines on what hurts.
  2. 02NDA and read accessWe sign the NDA first. Then you give me read access to the repository.
  3. 03Brief look reportScore, headline risks, quick wins. Free in the queue, or within 48 hours fast-tracked.
  4. 04Full auditOptional. Expanded findings, critical fixes, and a strategy.
  5. 05Hands-onOptional. I oversee the refactoring or lead the migration.

Under NDA. By default I review with AI tools (Claude Code, Codex) on my side. Prefer human-only, or AI inside your own environment? That's available with a paid full audit.

Pricing

EUR, excl. VAT. EU businesses with a VAT number pay 0% (reverse charge). Early-client rates. They'll change as the work speaks for itself.

Brief lookFreeScore, headline risks, quick wins. Joins the queue, handled in the order requests come in.
Brief look, fast-tracked€50Same report, fixed price, delivered within 48 hours.
Full audit€250About 5 hours. Detailed findings, critical fixes, infrastructure constraints, and a prioritized plan with a scoped proposal for next steps. Same price with AI tools on my side or AI inside your own environment.Full audit, human only€1,000About 20 hours. Same scope, but I read the code manually instead of using AI tools.
Hands-on work€50 / hourI do the fixes myself, or I advise your team: ticket planning, sprint support, reviews. Prepaid hour blocks (e.g. 20 hours = €1,000), used on an agreed weekly schedule and billed in full hours. Lower rate if we use your own AI infrastructure.

The full audit tells you exactly what to fix. After that, you decide whether to do it with your team, with me, or both.

Request a Tech Debt Check

Two minutes. I reply personally.

Who reads your code (all under NDA)
Review type
Only the manifest: no source, lockfile or .npmrc. Check it for private registry URLs before pasting. No package.json at hand? Leave it empty and describe the app below; a repository link helps.
Rough project size (lines of code)
Covered by NDA before any code changes hands.

By sending this you agree to the processing described in the privacy note. Protected by reCAPTCHA: Google Privacy Policy and Terms apply.

Questions

Is the brief look really free?

Yes. No card, no call, no obligation. Requests are handled in the order they come in. If you need it sooner, the fast track costs €50 and the report lands within 48 hours.

What do you need from me?

For the request: your package.json if you have one, and a few lines on what hurts. If we go ahead, we sign an NDA first, then you give me read access to the repository.

Do you use AI on my code?

By default, yes: Claude Code and Codex, on my own setup, under NDA. If your code must not leave your infrastructure, I can work with AI inside your environment (full audit, €250) or read the code by hand (full audit, €1,000).

Which stacks do you review?

JavaScript and TypeScript: React, Vue, Node, Express and NestJS, plus the build, test and lint tooling around them. Code written by people, by AI tools, or both.

How long does it take?

A fast-tracked brief look arrives within 48 hours. A free one depends on the queue shown at the top of the page. A full audit is about 5 hours of work, scheduled with you.

Can you fix what you find?

Yes, if you want. I can do the fixes myself or support your team with ticket planning, sprint support and reviews, at €50 per hour in prepaid blocks. The report is yours either way.