Contact us

Software project rescue

Software Project Rescue: Take Back a Stalled Build

Your developer stopped answering, or the agency missed another deadline and you cannot tell what is really done. First get the code and the accounts under your control. Then find out what you have, and decide whether to fix it or start again.

How the diagnostic works
  • Read-only access under NDA
  • Nothing changes in production without your decision
  • Keeping your code is always an option

What is software project rescue?

Software project rescue means taking back control of a build that stalled. Someone on your side recovers the code and accounts, stops further damage, finds out what actually works and decides what happens next. The aim is a clear decision: keep and fix what exists, or replace part of it.

The usual triggers are a freelancer who disappeared, an agency that keeps moving deadlines, or a handover that left nobody able to explain the code.

First 72 hours: what to do when a software project stalls

Act in this order. Get access to everything that runs the product. Freeze changes so nothing breaks while you look. Take an inventory of what exists and what works. Then get a verdict on what to do next. How fast each step goes depends on who holds the access today.

  1. Step 1

    Get access

    Code repository, hosting and cloud, domain, database, app store accounts and paid services. Ask for admin rights in your own name, not shared logins.

  2. Step 2

    Freeze changes

    No new releases until someone you trust has looked. Rotate passwords and keys the previous team held, once you have admin access.

  3. Step 3

    Take an inventory

    What exists, what runs in production, what was promised and what was delivered. Check the running product itself.

  4. Step 4

    Get a verdict

    A written assessment ranked by business risk: what to fix first, what is worth keeping and whether any part needs replacing.

What to get back from a developer or agency

Get back everything the product needs to run without them: the source code repository, hosting and cloud accounts, the domain and DNS, the database and backups, app store accounts, payment and other paid services, and the passwords and keys. Each should end up in an account registered to your company.

A checklist of what a founder needs back to own the product
AssetWhy it mattersWhat to ask for
Source code repositoryWithout it, nobody else can change the productAdmin access or transfer to your organisation, with full history
Hosting and cloudThe product runs and bills hereOwnership of the account, or admin access plus billing in your name
Domain and DNSControls your website and emailRegistrar account in your name
Database and backupsYour customers’ dataAdmin access and a recent backup you have checked
App store accountsNeeded to release mobile updatesApps under your developer account
Payment and paid servicesMoney flows and integrationsOwner access to each service and its API keys
Passwords, keys and documentsHidden dependencies break things laterA full list, then rotate everything

My web developer disappeared. What do I do now?

Start with what you control. List every account the product uses and check which ones are in your name. Ask the developer in writing for access to the rest, and keep the message factual. Pause any new releases. Then have an independent engineer review what exists before you hire someone to continue it.

  • Check the contract for clauses on code ownership, delivery and handover.
  • Look in your own email for invoices and sign-up confirmations. They show which services exist.
  • Do not let a new developer start changing the code before someone has taken stock.

The agency keeps missing deadlines. How do I know what is really done?

Judge the product you can run yourself. Ask for a demo of each promised feature on a test environment you can use yourself. Compare it with the milestones you paid for. An independent review of the code shows what is finished, what is half done and what was never started.

You do not have to end the contract to get an outside view. A diagnostic can run while the agency keeps working, with read-only access that changes nothing.

What do rescue reviews find in other people’s code?

The common pattern is code that compiles, deploys and looks fine while a critical flow is silently broken. Money that never returns, forms that discard data, access rules that show one customer’s data to another. These cases are published Valletta work, with clients anonymised.

Rewrite or keep fixing after a failed software project?

Keep fixing when the data model can carry the product and the problems are specific. Rewrite only when the data model cannot carry what the business needs, and someone can show why. A failed project feels like it needs a fresh start. Ask for the evidence before you choose.

In the four AI-built audits on our cleanup page the answer was refactor every time. Keeping your code and your developers is always an option in the report.

Who looks after the project while you decide?

An interim technical owner. They bring repositories, cloud and pipelines back under your control, write an assessment of what exists and prepare a runbook and handover for the permanent lead. Commercial terms with the outgoing vendor stay with you. The length of the engagement is agreed up front.

How the Fractional CTO team works

What does a project rescue look like in practice?

A German warehouse software company had a signed $180,000 annual contract, but billing could only start after the customer accepted the rollout. After two failed launches, our CTO narrowed the release to the workflows that acceptance needed and ran the next rollout with the warehouse team. The customer accepted it and paid the first $15,000 invoice within six weeks.

  • First $15,000 invoice paid within six weeks

    A stalled release became a $180,000 customer contract

    An 18-person German warehouse software company had a signed $180,000 annual contract, but billing could only start once the customer accepted the rollout at its first warehouse. After two failed launch attempts, our CTO joined a warehouse shift and narrowed the release to the workflows the customer needed for acceptance: receiving stock, picking orders and handling exceptions. He paused the broader backlog and ran the next rollout alongside the warehouse team. The customer accepted the release and paid the first $15,000 invoice within six weeks.

    Client name withheld.

Is your situation a little different?

After a rescue, founders usually face one of two next steps. A new vendor sends a quote for the remaining work, or an investor wants to know the state of the technology before a round. These pages cover both, and the hub explains the diagnostic and the team behind it.

Frequently asked questions

Yes. A CTO from our team helps you recover access, freezes changes and reviews what exists. You get a written verdict on what to fix first and what is worth keeping. Who continues the build is your decision: a new developer, your own hire or a team you choose.

It depends on your contract and the law where it applies. In many places the author keeps the rights unless a written agreement assigns them to you. Check your contract for an IP or ownership clause. This is general information, not legal advice. For a dispute, talk to a lawyer in your jurisdiction.

Usually keep fixing. Rewrite only when the data model cannot carry what the business needs. In the four AI-built products we audited, the answer was refactor every time. A written review ranked by business risk shows which parts are worth keeping before anyone proposes starting again.

We tell you exactly what to ask for and check what you receive is complete and runs. The request itself, and any commercial or legal terms with the developer, stay with you. If the code already sits in an account you own, the review can start from there.

Tell us where the project stands

Say who built it, what you can access today and what stopped. On a 30-minute call we tell you what to secure first and whether a diagnostic is the right next step.

Prefer to talk?

Let’s talk