AI built the prototype. I turn it into code that holds.
ChatGPT, Cursor, and Claude get you to a demo quickly. Then the real work begins: broken imports, missing architecture, unsafe secrets, unclear data flows. I find the breakpoints and repair the important ones – so you can decide whether the prototype is salvageable.
Technical audit and implementation, not legal advice. I check visible signals, integrations, and delivery issues; legal texts and binding legal assessments remain the work of lawyers or privacy consultants.
Typical symptoms
When AI code gets expensive
Build
Check
Architecture
Triage
Fix plan
Prioritized
Positioning
What makes this different
Senior repair, not prompt service
For projects where a demo exists but technical substance is missing. I work on structure, not the next quick answer.
Fixed prices for clarity
No open-ended timesheet. You buy a diagnosis, stability, and a viable next step.
Triage before rewrite
Do not rebuild everything. First we clarify what is salvageable, which errors are dangerous, and how the project becomes manageable again.
How it works
From repository to stable next step
- 1
Secure repository
I review structure, dependencies, build, runtime errors, and critical user flows.
- 2
Limit damage
Broken imports, wrong configuration, dead files, and fragile AI shortcuts become visible and prioritized.
- 3
Repair deliberately
Depending on the package, I stabilize core flows, architecture, tests, and deployment readiness.
No legal advice
I check technical signals and give you a clear priority list, not legal assessments.
Start only after approval
Booking creates the order. Work begins only after you confirm the start in the customer panel.
Clear priorities
You receive concrete findings sorted by risk, effort, and impact.
Documented results
Every package includes a written summary of findings and next steps.
Packages
Diagnosis first. Then targeted repair.
The prices are calculated as senior fix packages: enough scope for real analysis and repair, but small enough to create clarity before a complete rewrite.
AI Code Triage
When the project starts, but nobody knows why it keeps breaking.
Code review, build/import check, and rescue plan within two business days.
one-time payment, plus VAT where applicable
- Repository check for broken imports, missing packages, and build errors
- Assessment: repair, restructure, or discard
- Prioritized fix list with effort estimate
- No implementation included
Risk reversal
If the triage finds no actionable problems, we discuss how the price can still be used meaningfully.
Commercial terms
- one-time payment, plus VAT where applicable
- work starts only after explicit approval
- implementation scope is limited to the selected package
Stabilization Sprint
For prototypes that almost work but are not reliable yet.
Triage plus up to ten hours of targeted repair and stabilization.
one-time payment, plus VAT where applicable
- Repair build, routing, and core errors
- Turn unclear AI structure into understandable modules
- Clean up basic error handling and configuration
- Short technical handover with next steps
Risk reversal
If the analysis shows that production rescue is not the next sensible step, the amount is fully credited when you book a stabilization sprint.
Commercial terms
- one-time payment, plus VAT where applicable
- work starts only after explicit approval
- implementation scope is limited to the selected package
Production Rescue
When an AI prototype needs to become a real product.
Multi-day cleanup for architecture, security, tests, and deployment readiness.
one-time payment, plus VAT where applicable
- Clean up architecture and data flow
- Defuse security risks, secrets, and API errors
- Establish tests, typecheck, and build pipeline
- Deployment readiness and documentation for further development
Risk reversal
If the analysis shows that production rescue is not the next sensible step, the amount is fully credited when you book a stabilization sprint.
Commercial terms
- one-time payment, plus VAT where applicable
- work starts only after explicit approval
- implementation scope is limited to the selected package
Code Guidance
Ongoing senior control while you continue developing with AI.
Monthly allowance for reviews, small fixes, and technical guardrails.
monthly subscription, plus VAT where applicable
- Up to five hours of review or fixes per month
- Pull request or repository checks
- Validate architecture and tooling decisions
- Ideal for founders with ongoing AI prototyping
Commercial terms
- monthly subscription, plus VAT where applicable
- minimum term: 3 months
- unused monthly time does not carry over
- larger changes and emergency work are scoped separately
Results and timelines depend on the concrete state of your website or repository. You only book implementation if the priorities make sense.
FAQ
Common questions
Can every AI codebase be rescued?
No. Sometimes a controlled rebuild is cheaper. That is exactly what the triage clarifies: afterwards you know what can be repaired and what should be rebuilt.
Does the code need to be public?
No. A private repository is possible if secure access is set up. Detailed access instructions follow after booking.
Do I only get an assessment or also fixes?
The triage provides diagnosis and plan. The sprint packages include concrete repair time with documented next steps.
Can I keep developing with AI afterwards?
Yes, ideally with clearer guardrails so new AI changes cause less damage.
Stabilize instead of guessing
If you already invested time in the prototype, a hard technical triage is usually worth doing first.
Technical audit and implementation, not legal advice. I check visible signals, integrations, and delivery issues; legal texts and binding legal assessments remain the work of lawyers or privacy consultants.