DEVHERO Write on WhatsApp
HelplineReal humans, not AI. Help within 24h
Human First
legacy code nobody understands

Legacy code nobody understands is running your business. We fix it within 24 hours.

You depend on a system that predates the current team.

  • The last person who understood it has left
  • A small business change has no obvious place to go
  • Production behaviour disagrees with the comments

One free session per client: the first 4 hours with a specialist, plus unlimited AI as a tool.

A real senior specialist leads. Name on first reply. AI is a tool; a human checks the result.

01

The problem

What this problem looks like, how you can tell, and what is at stake if it stays.

What this problem looks like

What an unreadable legacy system does to a business, why teams freeze, and why mystery is not heritage.

A change takes weeks because nobody is sure

The people who built it left. Comments are missing or misleading. One module talks to a database nobody documented. A small change takes weeks because everyone is afraid of side effects. Tests, if they exist, do not cover the paths that make money.

The team is stuck and a little embarrassed

The team feels stuck and a bit embarrassed. New hires bounce off it. You keep postponing features because the foundation is a black box.

Software that earns money still needs a human who can change it

Software that earns money still needs a human who can change it. Mystery is not heritage.

You want a change, a fix, or a safe way forward — not a lecture on how the old code should have been written.

Legacy code nobody understands is still running invoices, bookings, or logistics. The problem is not that it is old.

Documentation, if it exists, describes a system that is no longer there. Tests, if they exist, skip the paths that make money.

Signs you have legacy code nobody on the team understands

These are operational signals, not aesthetic complaints about an old language.

The last person who understood it has left

Pull requests stall because nobody will approve a change in a module they cannot explain.

A small business change has no obvious place to go

A tax rule, a new rate, or a partner file format should be a bounded edit.

Production behaviour disagrees with the comments

The comments, or the wiki, describe a rule the running system no longer follows. Operators have private workarounds.

Builds and deploys are ritual, not a documented path

Only one machine, or one person, can ship. The steps live in a notebook, not in the repository.

Incidents are resolved by memory

When something breaks, the fix is 'restart this box' or 'run that script' with no written cause.

If nobody can change the system, and if a human can

Before — if you leave it

If you leave it

If no one understands the system, you cannot fix a bug quickly, you cannot pass it on, and you cannot add the feature a customer already paid for. Eventually you will be tempted by a full rewrite that you cannot afford to get wrong.

After — if we close it

If we close it

You get a map of the dangerous parts, a safer way to change what you asked for, and notes a future specialist can use. The system is still yours. It is just no longer a rumour.

Recognize your situation? Message us, we will sort it out.

02

Why it happens

Where this usually comes from, why it shows up in fast or AI-assisted work, and how to check it yourself.

Why undocumented legacy systems become unmaintainable

Age alone does not make a system unworkable. Missing ownership and missing characterisation do.

Knowledge left with a person, not with the repo

The design lived in someone's head, a private chat, or a machine that was never inventoried.

The system grew by exception

Each urgent fix added a special case. After enough exceptions, there is no single rule to read.

Tests never covered the money path

What was easy to test was tested. The ugly, valuable path was left as a manual ritual.

Integrations outlived their owners

A file drop, a queued message, or a nightly job still runs, and nobody can name the counterpart on the other side.

Rewrite talk replaced reading

The team spent its energy arguing about a future stack and never invested in a map of the present one.

Why AI tools and rushed teams struggle with legacy code

An assistant can summarise a file and still miss the stored procedure, the side-effect in a trigger, or the batch that runs at 2 a.m.

How to start understanding a legacy codebase yourself

You can do a first pass without a rewrite plan. The aim is a map, not a judgement of the original authors.

  • List the money paths and the batch jobs

    What must work on a weekday: invoicing, import, export, login, the nightly run.

  • Run the system and watch it, if you still can

    A running environment, even if fragile, tells you more than a frozen repo.

  • Follow one change that the business already wants

    Pick a real request and see how far you get.

  • Search for the hidden schedulers

    Cron, task schedulers, database jobs, and 'temporary' scripts in home directories are often the real application.

  • Identify the data you must not lose

    Backups, de-duplication, and a way to restore matter more than a new framework conversation.

Recognize your situation? Message us, we will sort it out.

03

How DevHero fixes it

A senior specialist leads. AI is a tool. A human checks the result.

Legacy is normal. Most lasting products have a core that grew under pressure. The problem is not that it is old. The problem is that no one now can stand for a change.

DevHero is Human First. A specialist with around 20 years of experience is used to reading unfamiliar systems and making a safe next step, not a fashion rewrite. AI can help search and summarise. A human decides what is load-bearing. An expert on payroll checks the result.

Around 20 years of experience is useful here because the work is judgement: what is load-bearing, what can be characterised with a test, and what should wait.

How a typical weekday case moves

  1. 01Message received
  2. 02Specialist named
  3. 03Diagnosis
  4. 04Fix
  5. 05Human check
  6. 06Handover

How to get a specialist into a legacy system

Write on WhatsApp or Telegram. We take hard, unreadable systems when we can help.

  1. Tell us what the system does for the business

    Name the product, the stack if you know it, and the change or incident that is blocked. A repo or a staging URL helps.

  2. A named specialist answers

    Within a minute you should see the name of the person who will lead. They are alerted, and can be called if they do not respond.

  3. A map and a safer change, in 24 hours when in scope

    On a typical weekday we aim to map the dangerous parts and make the change you actually need, or to tell you honestly if the next step is a larger rescue.

How DevHero works with legacy code nobody understands

We read the running system, characterise the risky path, and make a change a human will stand for. A rewrite is a last resort, not a first pose.

We read the running system, the data, and the deployment, not only the repo myth. We identify the seams, add characterisation where it is risky, and make the change you need — or say if the next step is a larger rescue. We do not start a greenfield rewrite in a day.

  1. Recover a way to run and observe it

    Access, environments, jobs, and data matter as much as the repo myth.

  2. Map the seams around the request you actually have

    We identify what is load-bearing for that change, not a complete encyclopaedia on day one.

  3. Add just enough characterisation

    A test, a script, or a recorded behaviour around the risky path, so the edit has a chance of being safe.

  4. Make the change, or recommend a larger rescue

    If the work clearly exceeds a session, we say so before we pretend otherwise.

  5. Leave notes a future specialist can use

    The system should be a little less of a rumour than when we found it.

What you get in the first 24 hours on a legacy rescue

For a typical weekday case in scope, the aim is a map and a safer next step — not a new product in a new language.

  1. A map of the parts that matter for your current request, including jobs and data if we can see them.

  2. A change, or a clear reason why that change is not safe yet.

  3. Notes on what we learned that the next person can use.

  4. A human-checked result, and an honest recommendation if a rewrite or a longer programme is the real work.

  5. A statement of what we did not have time to read.

What to send us for a legacy system nobody understands

Anything that helps us run or watch the system is useful. Do not send live secrets in the first message.

Repo, or a copy that actually builds

Plus any second repository that people forget: scripts, 'ops' folders, or a database project.

How it is supposed to run

A staging URL, a VM, a compose file, or the person who still knows the ritual.

The change or the incident you care about now

A business rule, a bug, or a launch that is blocked.

Jobs, integrations, and data stores

Cron, file drops, message queues, and the names of the databases.

Backup and restore status, if you know it

Especially if the system is the only copy of the data.

Quick and urgent fixes

First 4 hours with a specialist + unlimited AI: free, once per client. Further work is paid, on request. A human leads and checks the result.

Free (first session)

Complex projects

Larger work uses a different rate, on request. We say so before we continue.

On request

Weekdays, then weekends

The base plan is weekdays only. Weekend work is a separate product. Any surcharge is stated up front, before any call, and depends on complexity. Amount: on request.

On request

The first session is free: the first 4 hours with a specialist, plus unlimited AI, once per client. Work beyond that, the complex-project rate, and the weekend rate are paid, on request.

One free session per client: the first 4 hours with a specialist, plus unlimited AI as a tool.

Recognize your situation? Message us, we will sort it out.

04

Stay fixed

What not to do next, the stacks we see in this kind of work, and how to keep the fix in place.

What not to do with unmaintainable legacy code

Some popular moves spend a year and still leave you without a working path.

Starting a full rewrite before you can describe the current behaviour

You will re-implement folklore and miss the exceptions that keep the business alive.

Letting an assistant 'translate' the whole tree

Generated translations look complete and can still drop a trigger, a rounding rule, or a nightly side-effect.

Editing production directly because 'that is how it has always been done'

An emergency change should be rare and recorded, not the default workflow.

Pretending the wiki is the system

If production disagrees, production is the specification until you have characterised it.

Waiting for a quiet quarter that never comes

The map can start with the next change you already need.

Kinds of legacy stacks we are asked to read

Tell us the language, the database, and how it is hosted. If we should not take it, we will say so.

Older web and business systems

PHP, Java, .NET, Ruby, Python, Delphi, and similar applications that still take orders or keep records.

Databases that are also the application

Stored procedures, triggers, and large SQL Server or Oracle estates.

Batch and integration glue

Cron, file transfers, ETL jobs, and message queues that nobody has drawn.

Mixed estates

A modern front end talking to a core that predates the current team.

How to keep a legacy system from becoming a rumour again

After a rescue, a few habits stop knowledge from leaving with the next person.

Right after the fix

Write down the money paths and the jobs

A short, current note in the repo beats a faded wiki.

Characterise before you 'just quickly' edit a core module

A test or a recorded example is part of the change.

Make the deploy path something more than a ritual

If only one machine can ship, that is still a risk.

Inventory integrations twice a year

Names, owners, and what happens if they stop.

As you keep shipping

Do not let rewrite talk replace the map

A future stack is allowed. It is not a substitute for understanding.

Onboard people on the running system, not only the slides

They should see the job that makes money.

Recognize your situation? Message us, we will sort it out.

05

Questions about this problem

Straight answers about this problem, the session, and what happens next.

Will you rewrite the whole system in 24 hours?

Only if that is truly the right move, and not as a default promise.

What if there is no documentation?

That is a usual starting point. We reconstruct enough behaviour to work safely, and we write down what we learned.

Can you work with a very old language?

Tell us the stack in the first message.

Our core is stored procedures. Does that count?

Yes. A database that is also the application is a common legacy shape.

Should we freeze all changes until a rewrite?

A freeze can reduce risk for a short period, and it can also freeze the business.

How do I start?

Write on WhatsApp or Telegram about the system and the change you need. No form, no web chat.

Will you invent a 'legacy score'?

No. We will describe what we could read, what we changed, and what remains unknown.

Is this the same as technical debt blocking features?

They overlap. This page is for a system nobody understands.

Is the first session really free?

Yes. Each client gets one free session: the first 4 hours with a specialist, plus unlimited AI as a tool. The 24-hour line is how quickly we aim to start and close a typical weekday case in scope, not 24 hours of unpaid work. Further work is paid, on request.

Human First

Legacy code nobody understands is running your business. We fix it within 24 hours.

If a change is already blocked by code nobody understands, write on WhatsApp or Telegram. You will talk to a person.

One free session per client: the first 4 hours with a specialist, plus unlimited AI as a tool.

Human First

Choose a channel

No web form, no web chat. Write on WhatsApp or Telegram and a specialist replies. Your message is prefilled with the problem and the page.

Telegram opens the chat with @brnzai. It may not paste this text automatically — copy it and send it there: