Service 06 · Audit · Stabilise · Refactor
The back-end nobody wants to touch. We will.
Take over a system the original team has left, stop the fires, then make it fast and safe to change. Without a rewrite, unless a rewrite is honestly cheaper.
01 — What's included
Stop the fires. Then make it boring.
Boring is the goal: deploys nobody notices and pages nobody gets.
02 — Why not a rewrite
A rewrite is a bet. A refactor is a plan.
Most rewrites take twice as long as estimated and ship with the same bugs in a new language, because the old system was the only spec.
A refactor keeps the system running and the revenue flowing while it changes underneath. Each step is small enough to reverse and measured on deploy frequency, error rate and p95.
Sometimes a rewrite really is cheaper. Then the audit says so, with numbers, and we help you plan it.
Honest audit
If a rewrite is cheaper, the audit says so, with numbers
No downtime
Changes ship behind flags; old and new run side by side
Your team included
We pair with whoever will own it after us
Measured
Deploy frequency, error rate, p95: before and after
03 — How it runs
What you get, week by week.
Week 1
Audit
You get: a report with risks, a plan and an estimate. Yours to act on, with us or anyone.
Weeks 2–3
Stabilise
You get: monitoring, backups, the top fires out, a system that stops surprising you.
Weeks 4–12
Refactor
You get: module by module, tests first, a weekly report with the metrics.
Handover
Boring
You get: docs, runbooks, pairing sessions, and a support option if you want one.
04 — Pricing
Audit
Fixed price. A report with risks, plan and estimate, yours to act on with anyone.
Rescue team
Senior engineers on the system until it is boring. Cancel with 2 weeks' notice.
05 — Questions
About rescue projects.
That is the usual starting point. Access to the repo, the servers and whoever uses the system is enough. The audit reconstructs the rest.
Only if the numbers say so, and then with a plan that keeps the old system running until the new one is proven. Most systems do not need it.
Yes, and we prefer it. The handover is easier when the people who will own the system were in the room for the changes.
Node.js, Python, PHP and Ruby back-ends, React and older front-ends, PostgreSQL and MySQL, on most clouds. If it is something else, we say so on the first call.
Book an audit
Show us the repo. We tell you what we see in 5 days.
Read-only access and one call with whoever runs the system. The audit report is yours whatever you decide next.