Diagnosis and recovery plan
A map of the actual production configuration, the reproduced failure, a backup package and agreed acceptance checks.
Production recovery · One app
Get the app
working again.
For owners of a stalled Convex and Vercel app who need a controlled repair and evidence that the important path works.
Start with the URL, symptoms and stack. Keep passwords, tokens and customer data out of your initial email.
New fixed-scope pilot
$4,500 USD
One accepted small app. $2,250 after written scope acceptance; $2,250 after the agreed checks pass.
Three-business-day delivery target after complete access and an agreed maintenance window. Provider and access delays can extend that window.
One active recovery at a time. Your scope and start date are confirmed before any payment.
PreserveBack up data and files before changes.
RepairFix the agreed production failure.
ProveVerify the data and customer journey.
What you receive
A working release, with the evidence behind it.
Scoped repair and release
A reviewed change with focused tests, data and file comparisons, a controlled deployment and one verified customer journey.
Handoff and recovery instructions
A release record, compatible rollback or roll-forward instructions, remaining limits, a 30-minute handoff and a next-business-day smoke check.
Recent owner-operated work
Two apps restored.
Every original file accounted for.
A local directory and creator-profile app had unavailable backends. Recovery also required deployment, authentication and ownership-control repairs.
Across both, we preserved 242 original application records and 11 stored files. Record IDs and creation timestamps survived; file bytes matched the backups. Necessary content changes were recorded separately.
What was checked
- 35 focused behavioral tests for each repaired app
- Public pages and authenticated owner views
- Ownership boundaries and unauthorized writes in the tested paths
- Signed authentication-webhook delivery and an internal email delivery
- Matching frontend/backend configuration and stored-image URLs
These were our own apps, not paid client engagements. No real purchase or customer booking was submitted, and no revenue increase was measured. One Google OAuth path remained untested and one preserved RAW image remained browser-incompatible. This was bounded repair, not a complete security audit.
Fit for this pilot
One existing Next.js/TypeScript repository, one Vercel project and one Convex application, with recoverable source and backups. Existing Clerk authentication; one Stripe or Clerk Billing source, and one existing Resend sender if needed for the selected journey.
Pilot scope: up to 40 application tables, 10,000 documents, 100 stored files and 250 MB of stored-file data. These are booking limits. Larger or continuously active datasets need separate scoping.
You provide authorized access, a controlled test identity and an agreed maintenance window. Provider charges and any required account resolution remain yours.
A defined recovery
The preflight identifies the failure, intended changes and acceptance checks before the job is accepted. An unexplained data change or failed agreed check remains incomplete.
The pilot covers one recovery and customer journey. Rewrites, new features, database merges, unavailable historical files, full security audits, 24/7 response and ongoing maintenance are outside this service.
We use supported provider operations and adequate authorized capacity. We do not promise zero downtime, automatic zero data loss, reduced hosting costs or account reinstatement.
Start with the failure
Send the URL and what stopped working.
I’ll review fit, define the acceptance checks and confirm the scope before you commit.
Request a fit review ↗