Software project rescue for builds that stalled and agencies that stopped answering
Half-built, over budget, and whoever was building it is gone or going. We take custody of what exists, tell you honestly whether it can be finished or has to be restarted, and then do that. No blame, no rewrite reflex.
- You get the assessment even if we do not take the build
- Everything transfers into your accounts, not ours
- Fixed price after the assessment, billed across milestones
Why this is worth building, and why us.
Rescues arrive in a familiar state. There is a repository nobody can build, an invoice history that does not match the progress, a developer who has gone quiet, and a founder deciding whether to keep paying. The technical problem is usually the smaller half. The larger half is that nobody can tell you what you actually have.
That is the first thing we fix. Before any code gets written, we get it building on a fresh machine, read what is there, and produce a plain-language account of how complete it really is. Everything else, whether to finish it, rebuild it, or stop, is a decision you can only make once that account exists.
What a rescue actually involves
- 01Getting it to build and deploy from a clean checkout, which is often where the first day goes
- 02An honest completion estimate: what works, what is scaffolding, what was never started
- 03Recovering access: repositories, domains, hosting, databases, and the accounts they live in
- 04Finding the data, and making sure there is a backup of it before anything else happens
- 05A written call on finish versus rebuild, with the cost of each
- 06Then doing the work, with a demo every week so you can watch it move
Priced by what is left standing
The only honest first step is an assessment, because a project that is eighty percent done and one that was described as eighty percent done look identical from the outside. Three days of reading gives you a real number for everything after it.
We take what you have, get it running, and read all of it. You get a written verdict: how much is genuinely done, where the risks are, whether finishing is cheaper than restarting, and a fixed price for each path. Yours to keep whatever you decide, and credited back in full against any work you start within 30 days.
- Takeover$3,4973 weeks
Getting custody and control back before deciding anything bigger.
- A clean build, a working deploy, and documentation of how both happen
- All accounts, repositories, and infrastructure moved into your name
- The data audited and backed up somewhere you control
- A security review of what the previous team left exposed
- A prioritised backlog of what remains, estimated honestly
- Most chosenFinish the Build$7,9978 weeks
The work is salvageable and it needs to reach the launch it never got to.
- Everything in Takeover, plus the remaining product built out
- The half-finished parts either completed or removed, not left ambiguous
- Tests around the flows that matter, so the next change is not a gamble
- A weekly demo, so progress is something you watch rather than something you are told
- Launch, and thirty days of support after it
- Rebuild$14,99710–12 weeks
The assessment says the foundation cannot carry what you need.
- A fresh build on Next.js, TypeScript, and Postgres, keeping the product decisions you already paid for
- Your existing data migrated, verified, and reconciled
- The features that were actually used, not the full original scope
- Full handover: code, infrastructure, docs, and a walkthrough
- The old system kept running until the new one has earned the traffic
- You get the assessment even if we do not take the build
- Everything transfers into your accounts, not ours
- Fixed price after the assessment, billed across milestones
- We will say when finishing beats rebuilding, including when that means less work for us
Rescues are the one place we refuse to quote blind, because the gap between the state a project is described in and the state it is in is the entire problem. The $497 buys a real answer, and often enough that answer saves you from a rebuild you were about to be sold. If we think the honest move is to stop, we will say that too.
Why the first question is not technical
The instinct in a stalled project is to ask what is wrong with the code. The more useful question is what you actually own. We regularly find domains registered to a former contractor, hosting on a personal account, a database with no backup, and a repository one person holds the keys to. None of that is a coding problem, and all of it is a business risk that gets worse the longer it sits.
So a rescue starts with custody. Accounts into your name, code somewhere you control, data backed up. That work is unglamorous and it is the part that stops the situation deteriorating while you decide what to do next.
Finish or rebuild, decided by evidence
Every agency that inherits someone else’s project wants to rebuild it. It is more work, it is more comfortable, and it is easy to justify. Sometimes it is even right. But it is a conclusion that should come from reading the code, not from the awkwardness of working in someone else’s.
The things that genuinely force a rebuild are narrow: a data model that cannot represent what the product now needs, a framework that is out of support, or a codebase with no coherent structure at all. Ugly code is not one of them. Ugly code that works and is understood is cheaper to finish than a clean rewrite is to start.
What went wrong last time, so it does not again
Stalled projects usually share a mechanism rather than a villain. Scope that grew without the price or the date moving. A fixed bid that stopped being profitable, after which the work quietly deprioritised. Progress reported in percentages nobody could verify. A single developer who became the only person who understood it.
We work against each of those directly: fixed scope agreed in writing, weekly demos of running software rather than status emails, everything in your accounts from the first day, and documentation written as we go, so the project never depends on one person staying interested.
Not claims. Shipped work.
- 01ReadyShortlistA marketplace with multiple user types built end to end, including the operational tooling around it.See the work
- 02Altamar ChartersA booking product shipped and running for a real business.See the work
- 03ChatberryA product taken from idea to production, including the decisions that shaped it.See the work
- 04Vibe coding cleanupIf the half-built thing was built with AI rather than an agency, start here instead.See the work
Common questions.
01What does a software project rescue cost?
$497 for a three-day assessment, which is the only part we can price before reading the code. After that it is $3,497 to take over custody and get control back, $7,997 to finish an eight-week build, or $14,997 for a full rebuild. The assessment tells you which of those you are actually looking at, and comes off the invoice if you continue with us.
02We do not have the source code. Can you still help?
Sometimes, and it is worth asking before assuming otherwise. Deployed applications can often be partly recovered, hosting accounts can be reclaimed with proof of ownership, and databases usually survive even when the developer relationship has not. If none of it is recoverable we will tell you quickly rather than bill you for looking, and what you learned about the product still shortens a rebuild.
03Will you tell us our developer did a bad job?
We will tell you what the code is like, which is a different thing. Plenty of stalled projects have competent work in them and failed on scope or money instead. If the work is genuinely poor we will say so plainly, because you need that to decide, but we are not interested in building a case against anyone. The report is written for your decision, not for a dispute.
04How fast can you start?
The assessment usually starts within a week and takes three days. That is deliberate: the situation tends to get worse while it waits, especially where accounts or data sit with someone who has stopped replying. Larger builds we schedule properly, but custody and backup we treat as urgent.
05Can you work with our existing developer?
Yes, and it is often the cheapest outcome. Sometimes what a project needs is a senior engineer setting direction and reviewing the work rather than a whole new team. If the assessment says that, we will offer it, and our fractional CTO engagement covers exactly that shape.
Get an honest read before you spend anything else
Tell us what was being built, who was building it, and where it stopped. The three-day assessment gives you a real answer on finish versus rebuild, and it is yours to keep whichever way you go.
- A real person replies, usually same day
- No sales sequence, no automated nurture
- Rescue Assessment is credited back if you go ahead