What happens to your system
if something happens to us.
Your system runs on hosting accounts in your company's name, from source code sitting in your own repository, over a database in Australia that you can copy any day without asking us. If Praxa Tech stopped existing tonight, nothing would stop running tomorrow.
The proof, and the limits, in public, before you ask.
Everyone asks these. They're right to.
So the answers are here, in public, where you can check them against what we actually do.
They come up in nearly every first conversation we have, usually near the end, and often with an apology in front of them. There shouldn't be one. You're deciding whether to run your operation on a system we build, and asking who holds the account is diligence, not rudeness.
You're a new company. What happens to the system we've just paid for if the person who built it is hit by a bus, or the company folds, or you decide this isn't worth doing anymore. Can we get our data out. Where does it actually live. Who else can see it.
They're the right questions, and often they aren't even the buyer's own fear. They're what their board will ask them, or their co-owner, or the risk register they have to fill in. They don't need a reassuring conversation they have to relay from memory afterwards. They need something they can forward.
The answer isn't that we won't fail. Any vendor can fail, and the large ones fail too, just more slowly and with more notice. The answer is that the arrangement is built so our failing doesn't become your problem.
Most software arrangements fail at exactly one point: the vendor holds the account.
What you own, and how to check it today.
Everything else follows from who holds the account. So we don't hold it.
| What | Whose name it's in |
|---|---|
| Cloudflare the application and the database |
Yours. The account is in your business's name, and your administrator can revoke our access without asking us and without telling us first. |
| GitHub the source code and its full history |
Yours. Your organisation, your ownership. Code lands there as it's written, not at handover. |
| Microsoft 365 logins and documents |
Yours already. We work inside your tenancy. We don't create a parallel one. |
| Xero or MYOB the accounting connection |
Yours. You authorise the connection and you can revoke it in your own accounting software without our help. |
Don't take our word for it. Go and look.
Ask us for the account details. Log in. Check whose name is on the account and whose name is on the billing. It takes four minutes and it's the single most useful thing you can do before signing anything with us.
Then ask the same question of every other vendor you're comparing us against. The answer will not be your name.
The code
On payment for each stage, the intellectual property in everything we built for you becomes yours outright. You can use it, change it, copy it and hand it to another developer, including one who competes with us.
The source code sits in your own GitHub organisation from the first commit, with its full history. There's no handover event where it moves from us to you, because it was never with us.
What we keep, and why it can't hurt you
Two things we're precise about, because a lawyer will ask. We keep ownership of our own general purpose tooling that existed before your project, and you get a perpetual, transferable, royalty free licence to it that survives us ceasing to trade. And your system uses open source components that stay under their own licences.
Nothing we keep can switch your system off, and nothing we keep can stop you leaving. That's the test to apply, and it's the one worth applying to any vendor.
Four situations, most likely first.
What actually happens on your side, in each one.
| Situation | What happens to you |
|---|---|
| Your developer is unavailable illness, leave, a bad week |
Your account manager tells you inside one business day and gives you a date. Hosting and the platform itself are unaffected. Development pauses until your developer is back. Your account manager runs the relationship, not the servers, and we're not going to tell you otherwise. |
| Praxa Tech stops trading | Your system keeps running. It's on your Cloudflare account, from your repository. Nothing needs to happen on day one. |
| There's nobody left to call | Identical to the row above from where you sit, because none of your access runs through us. You already hold the accounts and you already hold the code. |
| You want to leave | Tell us in writing. Within 10 business days we hand over everything, produce a full structured export of your database, and give your incoming developer a two hour handover call at no charge. No exit fee, no extraction charge, no notice period, and we don't slow walk it. |
The honest limit of "nothing stops"
If we vanish, your system keeps running. It also stops being maintained, and we'd rather you heard that from us than found it out.
Software dependencies get security patches roughly monthly, and some credentials expire on fixed dates. Our assessment: your platform runs safely unattended for about three months, runs with slowly growing risk out to about twelve, and shouldn't be left unmaintained past that.
A vendor who tells you it'll just keep running forever either hasn't thought about it or is hoping you won't.
There's a notice period here and it runs one way, on purpose. You can stop at any time with no notice and no fee. We'll give you 90 days' written notice before we ever stop providing hosting, monitoring or support for any reason other than non payment, and we'll hand over properly inside it. We should carry the notice risk. You shouldn't.
We are deliberately replaceable. Not out of generosity. A buyer who has to bet the operation on our survival shouldn't be buying from us, and we'd rather build the arrangement that way than argue about it later.
Any entry plan needs an exit plan.
Two ways out, and only one of them involves us.
Any screen, to Excel
Every list and report exports to CSV or Excel from the interface. It's the fastest route and it's the one your team will actually use, which matters more than the one that looks best in a document.
The whole database, direct
Your administrator can take a complete dump straight from your Cloudflare account using Cloudflare's own command line tool. One command, standard SQL out.
We'll run it with you once at go-live, so you've done it yourself rather than read about it.
One thing worth knowing now rather than in year three. The database is Cloudflare D1, which has no ODBC or JDBC driver, so you can't point Power BI or Excel straight at it. If live reporting against your own data matters to you, say so during scoping and we'll set up a nightly replica in a conventional SQL database you can connect anything to.
Your accounting never depends on us. Xero or MYOB keeps the books, and their whole job is keeping up with the ATO. That's the compliance heavy layer and we deliberately leave it alone. Your BAS, your books and your audit trail survive us completely. What sits with us is operational: stock, jobs, serial numbers, history.
We don't build payroll. Awards, EBAs and Single Touch Payroll change constantly, and getting them wrong is a wage underpayment problem, not a software problem. We'll integrate with a payroll vendor instead, for the same reason we leave accounting alone.
Behind a login, not behind obscurity.
Security through obscurity isn't security. Here's what's actually holding the door.
On your side
- Every login goes through your Microsoft account. Your multi factor authentication, your rules, your offboarding. Disable someone in Microsoft on a Friday afternoon and they lose the platform at the same moment.
- Permissions by role. The storeman sees the storeman's screens. Costs and margins are visible to the people you decide should see them.
- Encrypted moving and encrypted at rest. TLS 1.3 in transit. At rest, encrypted by Cloudflare under Cloudflare managed keys, which protects against a stolen disk rather than against Cloudflare itself.
- We never see your accounting password. The Xero and MYOB connections use a token you authorise and can revoke yourself.
- Internet facing, on purpose. Your team reaches it from the workshop, from home and from a phone, protected by authentication on every route. If you want it restricted to your office network or your enrolled devices, that's a configuration change, not a rebuild.
If something goes wrong, you hear it from us
If we become aware of actual or suspected unauthorised access to your data, we notify your nominated contact within 24 hours, in writing, with what we know and what we don't yet.
Daily updates until it's closed, logs preserved, a written incident report within 10 business days, and no public statement about an incident affecting you without your agreement.
We handle your data in line with the Australian Privacy Principles as though we were bound by them. Under the Notifiable Data Breaches scheme the obligation to notify sits with you, and you have 30 days to assess. Our job is to make sure you're never the last to know.
Where your data lives, stated precisely
Your database sits in Cloudflare's Oceania region, which for our clients means Australian infrastructure, and we disable the settings that would create read copies elsewhere.
Two things we won't blur. Cloudflare documents regional placement as a strong preference rather than a contractual guarantee, and the only hard residency guarantees they sell today cover the EU and US Government. And Cloudflare, Microsoft and GitHub are all US controlled companies, so US law can in principle reach data they hold no matter where it physically sits. Data residency is not data sovereignty, and anyone telling you otherwise is selling.
If your head contract flows down a sovereignty requirement, send us the clause. We'll tell you within five business days whether we can meet it, what it would cost, or that you should go elsewhere. We'd rather lose the work than have you fail an audit.
Nothing here is a surprise waiting for you in month six.
| What we don't have | What we do instead | What it means for you |
|---|---|---|
| Overnight support | Faults raised overnight are actioned from 8:30am. Two hour response when nobody can work. | Fine if you manufacture on Australian hours. Wrong supplier if you run 24/7. |
| ISO 27001, SOC 2 or IRAP | Microsoft single sign-on with your own MFA, role based permissions, encryption, Australian data residency. | Enough for most manufacturers. Not enough for a Defence flow down that names a scheme. Send us the clause and we'll answer in five business days. |
| A long trading history | Registered 2026. The experience behind it is a career in IT and development, and five years selling and implementing ERP systems. Live clients you can phone. | Judge us on the reference call and on stage one, not on the incorporation date. |
| Compliance updates spread across thousands of customers | Accounting and tax stay in Xero or MYOB. We don't build payroll. | The compliance heavy layer never depends on us. Operational changes on your side of that line are quoted in writing first. |
| A performance guarantee before you've seen it | Stage one. Fixed price. In your hands. You decide before stage two. | You judge a working system, not a promise. |
What if you can't finish?
The fair version of the bus question, and it deserves a straight answer. We stage the work so you're never exposed to an unfinished whole. Each stage is priced, scoped in writing, and delivered as something you can use on its own. You pay per stage and you decide whether the next one goes ahead.
We don't start a stage we aren't funded to finish, and we don't run more concurrent builds than we can staff. And if we've under quoted, you'll hear it from us rather than quietly receiving less. One uncomfortable conversation beats three quarters of a system.
Including the ones with an apology in front of them.
What happens if there's nobody left to call?
Your system keeps running. It runs on a Cloudflare account in your business's name, from source code in your own GitHub organisation, and none of your access runs through us.
Nothing needs to happen on day one. You'd need a developer eventually, and the honest window for that is three months comfortably, twelve at the outside.
Can you lock us out of our own system?
The accounts are in your business's name, and your administrator can revoke our access without asking us and without telling us first. You don't need our cooperation to remove us.
We ask one thing only, and it's for your benefit: tell us the same day, so we're not still deploying to a system we can no longer see.
Is there a source code escrow arrangement?
No, because escrow exists to hold code the vendor keeps, and you already hold yours. The source code sits in your own GitHub organisation with its full history, from the first commit rather than at handover. There's nothing for a third party to hold on your behalf.
It's worth asking any vendor who offers escrow what exactly is being escrowed and why they're holding it in the first place.
Where does our data actually live?
Your database sits in Cloudflare's Oceania region, which for our clients means Australian infrastructure, and we disable the settings that would create read copies elsewhere.
Two things we won't blur. Cloudflare documents regional placement as a strong preference rather than a contractual guarantee. And Cloudflare, Microsoft and GitHub are all US controlled companies, so US law can in principle reach data they hold no matter where it physically sits. Data residency is not data sovereignty, and anyone telling you otherwise is selling.
What does it cost us to run this without you?
Cloudflare bills the hosting. Beyond that, budget 40 to 80 hours for a new developer to get comfortable with the system, once, at market rates.
That's the number to plan for. Not a rebuild, which is what most people assume and what most vendors quietly rely on them assuming.
How long can the system run with nobody maintaining it?
Our assessment: safely unattended for about three months, with slowly growing risk out to about twelve, and it shouldn't be left unmaintained past that. Software dependencies get security patches roughly monthly and some credentials expire on fixed dates.
A vendor who tells you it'll just keep running forever either hasn't thought about it or is hoping you won't.
What if we want to leave?
Tell us in writing. Within 10 business days we hand over everything, produce a full structured export of your database, and give your incoming developer a two hour handover call at no charge.
No exit fee, no extraction charge, no notice period, and we don't slow walk it. An exit that goes badly is worse for us than it is for you.
Do you hold ISO 27001, SOC 2 or IRAP?
No, and we're not going to imply otherwise. What you get instead is Microsoft single sign-on with your own multi factor authentication, role based permissions, encryption in transit and at rest, and Australian data residency.
That's enough for most manufacturers. It isn't enough for a Defence flow down that names a scheme. Send us the clause and we'll tell you within five business days whether we can meet it, what it would cost, or that you should go elsewhere.
What if the person who built our system is unavailable?
Your account manager tells you inside one business day and gives you a date. Hosting and the platform itself are unaffected. Development pauses until your developer is back.
Your account manager runs the relationship, not the servers, and we're not going to tell you otherwise.
Ask us the hard question on the call.
Bring the one you'd normally save until after the contract. That's the conversation we'd rather have first, and if the answer is that we're not the right fit, we'd rather find that out now too.