Paste your GitHub, GitLab, or Bitbucket URL — or upload a ZIP straight from your machine.
We traverse every file, understand your stack, map dependencies, and identify the key moving parts.
Plain English. What each file does, why it exists, what breaks if you touch it, and how it all connects.
Download your owner's manual as a PDF. Read it like a book. Maintain your app like you built it yourself.
Every component, route, and function explained in plain language. No jargon. Written for the person who shipped it, not who wrote it.
A written map of how your app is structured — frontend, backend, database, APIs. See the whole car before you open the hood.
Flagged areas that need attention. How to add features safely. What to avoid touching. Think of it as the service schedule for your codebase.
Not just what your code does — but why. We surface the patterns your AI used so you actually level up while you read.
Your source code is never stored, logged, or used for training. Processed ephemerally. Gone the moment your manual is ready.
Beautifully typeset. Indexed. Sectioned. The kind of document you could actually put on a shelf. Because some things are better offline.
Started free. Noble. Unsustainable.
Real codebase analysis takes reliable processing, stable hosting, and enough compute to stop scans from timing out, falling over, or returning vibes in JSON.
So here is the deal: the scan is free, forever. The full manual is pay-what-you-want — name a number you think it's worth, and that unlocks the deeper analysis, the exports, and the maintenance guides.
No growth-hack wizardry. No fake generosity. No dark-pattern checkout that "accidentally" subscribes you. Just a product with server bills, an honour-system price tag, and the decency to keep working.
// What you actually get
Architecture, risks, maintenance guides, env vars, AI handoff prompts, glossary, service schedule and more.
Beautifully typeset. Shareable. Printable. Attached directly — no login required to open it.
Your manual stays available. Re-download any time in the next 30 days.
Public URL not required. Download your repo as a ZIP and drop it in.
Free Scan
$0
"What did I build?"
✓ Up to 5 MB / 15 files
✓ Quick Start + Identity Plate
✓ In-browser preview
✓ Public repos only
✗ No PDF download
✗ No email delivery
✗ No full analysis
Owner's Manual
PWYW
Name your price · one-time · "How does it actually work?"
✓ Up to 25 MB / 100 files
✓ Full 15-section analysis
✓ PDF typeset + delivered by email
✓ 30-day download link
✓ Public repos + ZIP upload
✓ Risk register + warning lights
✓ AI agent handoff prompts
What's in the full Owner's Manual?
15 sections: Quick Start, Architecture Overview, Key Systems, File Breakdown, Safe-To-Touch Map, Warning Lights, Risk Register, Running Costs, Environment Variables, AI Agent Handoff Prompts, Service Schedule, Troubleshooting, Known Unknowns, Glossary, and Next 5 Actions.
How much does it cost?
The scan is free. The full manual is pay-what-you-want — you name the price. Pay once, get the manual, done. No subscription, no auto-renewal, no "your trial is ending" emails. Skint students and funded founders pay on the same honour system. Push new code and want a fresh manual? Name a price for the new version.
Does it work on private repos?
Yes — via ZIP upload. Download your repo from GitHub/GitLab as a ZIP file and drop it into the uploader. No OAuth, no token, no access required. Your code is processed ephemerally and never stored.
How long does analysis take?
Free scans typically complete in 30–60 seconds. Paid full analysis runs in 1–3 minutes depending on repo size. You'll get an email when it's ready — no need to keep the tab open.
What happens after I pay?
Payment clears through an encrypted, PCI-compliant checkout — we never see or touch your card details. Then the real work starts: we run the full analysis pipeline, typeset your PDF, and email it to you with the file attached, usually within 2 minutes. A 30-day download link rides along in the same email.
How do I know my code is safe?
Because we built this the way we'd want our own repos handled. Your code is read in memory, analysed once, and dropped — never written to a database, never logged, never fed to a training set. No background access to your GitHub, no tokens sitting in someone's env file. We've shipped enough production code to be paranoid on your behalf.