Frequently asked questions

No. It runs in writing from start to finish, so every decision stays on record instead of depending on what someone remembers. There is one route in and it starts with the form on the How to order page. No call is required at any point.

Read the full version on How to order →

Every package starts with the form, including the ones with a fixed price on the service pages. I check the scope actually fits the package before anything is invoiced, rather than taking your money first and finding out afterwards.

Read the full version on How to order →

One document within 1 to 2 business days: scope, the full price, delivery dates, any monthly running cost, and a plain list of what is not included. The whole number goes in front of you once, and nothing gets added to it later.

Read the full version on How to order →

An invoice in English, priced in USD, with the transfer details on it, paid by international bank transfer the way most companies already pay contractors abroad. Your bank's outgoing transfer fee is yours, which is the normal arrangement. A receipt is sent by email as soon as the money lands. I do not take card payments today, and I would rather say so than let you find out at checkout.

Read the full version on Payment and handover →

Under $2,500 is paid in full up front, because splitting it costs us both more in admin than it protects. From $2,500 up it is 40% to start and 60% on handover, with the second invoice going out after you have the working system and have checked it. Large builds are split by phase, each phase with its own number and its own delivery date, and you can stop after any phase. You never pay ahead for work that does not exist yet.

Read the full version on Payment and handover →

While the scope is still moving, changes are free: cut something, add something, or ask why a line costs what it costs. Nothing is committed and nothing is charged at that point. After handover, bug fixes on anything inside the agreed scope are free, however many times it takes, for the full 60-day support period. New features or scope changes after the scope is agreed are quoted before I start, never after.

Read the full version on Support and refunds →

You do. The repository lives in your own GitHub account from the first day of work, not at the end, so you are holding the work the entire time and can hand it to somebody else whenever you want. Documentation is in English: a user guide with screenshots, a README your own developer can follow, and English comments in the code.

Read the full version on Payment and handover →

60 days from the day you accept the work. It covers bug fixes inside the agreed scope free however many times it takes, a reply within one business day, delivery on the date written down before the work starts, a written contract in English, full source code and documentation in English, and help with setup and go-live problems while you are first running it. I do not promise business results like sales numbers or Google rankings, because nobody controls those.

Read the full version on Support and refunds →

Cancel any time before I begin and you get everything back, no reason needed, no notice period and no cancellation fee. Once I have started, the time is spent and cannot be recovered, so there is no refund on work already done, but I do not invoice the remaining stages and you keep everything built so far including the source code. If I am the one who cannot continue, unworked stages are refunded with nothing held back as a handling fee.

Read the full version on Support and refunds →

These are listed on your scope document so none of them appear later as a surprise: domain registration roughly $15–25 a year, always in your name and not mine; hosting or servers depending on the size of the system, and plenty of small projects genuinely run on free tiers which I will tell you about; AI usage roughly $10–60 a month based on real usage for projects that use it; and a one-time $25 Google Play developer account for mobile app projects.

Read the full version on Payment and handover →

No. I am a registered sole trader, not a limited company, so the contract is with me by name. My commercial registration number is 1719900398604, issued by the Thai Ministry of Commerce, and my national ID number goes on the contract so you can check that the person signing is the person doing the work. There is no team and no subcontractors: the person you email is the person writing the code.

Read the full version on How to order →

I work with clients worldwide with one exception. I cannot supply software development, IT consulting or security testing to anyone located in Russia, Belarus, Cuba, Iran, North Korea, Syria, or the occupied regions of Ukraine. That is a legal restriction on this kind of service, not a preference, and it applies no matter who is paying.

Read the full version on Payment and handover →

Written permission from the owner. I only test systems you can show me written permission for, and the permitted scope goes into the contract before I touch anything. No exceptions, including when the site is obviously yours. It protects both of us.

Read the full version on Payment and handover →

No, and this is where it differs from a bot wired straight into an AI model. It searches your material first. If it finds nothing, it tells the customer it is not sure and passes the chat to your team with the original message attached. You are not chasing down wrong answers a week later.

PDF, Word and Excel files, photos of a menu or price list, text you type out yourself, even old chat transcripts you copy over. The closer the wording is to how your customers actually write, the better the answers come out.

Yes, that is the Connected Assistant package. The bot pulls live data from your system to answer, so it can say which slots are open, how many units are left, or where a parcel is. If your system is more unusual than most, I look at it and quote that part separately before any work starts.

A business with 30 to 80 chats a day usually lands around $10 to $25 a month. A few hundred chats a day takes it closer to $45. I work the number out from your real volume before you decide, and it is billed by the AI provider directly, not through me.

The two larger packages include an admin console, so your staff change answers themselves without waiting for me. On the smallest package you send edits to me during the 60 days. Either way the README, the code comments and the technical notes are in English, so your own developers can pick the project up without me in the loop.

You are hiring one developer directly. I work alone from Thailand, there is no agency behind me and nobody in between. I write back inside the hours set in the contract, worked out in your time zone, normally within 1 to 2 business days. Everything can run in writing, which I prefer anyway because it leaves a record of every decision.

I will not quote a number out of thin air. Send 10 to 20 of your real documents, I run them through and tell you what I actually got, field by field. If the result is not worth what you would pay, I will say so and tell you not to build it.

The system runs inside your own cloud account, so the data does not sit with me. If you want it, original files are deleted automatically once they have been read. Security work is my main line of work, so I build it that way by habit.

Every document gets a confidence score. Anything the system is unsure about, or where the totals do not match, is pushed into a review queue for a person. The clean, confident ones pass through. Your team checks the few that need it instead of all of them.

Yes. It can pull the tax registration number, document number, net amount and tax as separate fields. I did real bookkeeping before I wrote software, so I know which fields an accounts team actually needs and which ones are decoration.

Yes. Once the documents are data, reporting is the easy part. Connected Workflow already includes a spend dashboard, and anything beyond that is a small piece of extra work on top of what is already there.

No. I work in writing, in English: written scope, written updates, written handover docs. I reply inside stated working hours and I tell you what they are in your time zone. If your team can only make decisions on live calls, I am probably not the right fit for you.

They are different things. Off-the-shelf software is what you use behind the counter; your customers cannot book on it themselves. This is a system on the web that your customers use directly, at two in the morning if that is when they decide. If what you need is a till, buy the off-the-shelf one and I will tell you so. If you want customers booking themselves, it has to be something like this.

Yes. Everything listed in a package is covered by the price shown, with nothing added along the way. Anything not on the list I quote before doing, and you choose. The exclusions are written on this page, not buried in a contract.

Nothing recurring to me. You pay once and the system is yours. Hosting and a domain run about $40-70 a year, paid directly to the provider in your name, plus slip checking on the deposit package. If you would rather I looked after it, that is a separate monthly care plan and it is optional.

The first two packages are built for businesses that sell time and appointments, which covers salons, spas, clinics, workshops and most service trades. If your way of working genuinely does not fit, that is the third package, and I will tell you which one you are in at the first conversation rather than letting you buy one that falls short.

Yes. The code is yours, in your own account from day one, with documentation someone else can pick up. I do not lock clients in by hiding how it works.

You can, but I would not recommend it. The cost of extending a system depends almost entirely on the state of the code. Quoting without seeing it means pricing in risk, which means you pay more than the job is worth. Checking first makes the quote match the work, and if you hire me for the work I credit the check in full.

I will check it and give you a straight answer on whether upgrading or rewriting is cheaper. PHP 5 stopped getting security support in 2019, so staying on it is a risk that grows every month, but tearing it down can cost more than it needs to. Neither of us can answer that honestly without looking at the code.

Yes, and I prefer it. I write documentation and tests so your team can carry on without me. The goal is that you are not dependent on one outside person indefinitely.

I will answer that straight. My main stack is TypeScript and Python. For Laravel I would rather be judged on something you can inspect than on a number of years you cannot verify. GradeLens is a school grading system I built in Laravel 13, with the full source public and 26 tests covering the grade boundaries and the access rules, and it is running live where you can use it. It is on my Work page. If you read the code and it does not meet your bar, judge me on that.

That is exactly why I put tests around everything I touch and hand over in milestones you review before I continue. If something I did causes a problem, I fix it at no charge inside the cover period.

Because it is a different job. At that price you get the files cleaned so the site works again, which plenty of people can do and which is fairly priced for what it is. My work is finding how they got in and closing that route, which means reading server logs and the injected code rather than replacing files. If all you want is the site working again quickly, say so and I will scope it down, or tell you straight that a cheaper service is enough for your case.

Yes. For urgent cases I start looking before we settle the price. Two things help right now: do not delete anything, because deleted traces are deleted evidence, and ask your host to preserve the access logs. Many keep them only a few days.

Every core file is hashed against the official release for that exact version. A file that does not match was modified; a file that should not exist is foreign. The tool I use for that is one I wrote, and the source is public, called RaidPress, on my Work page. I will also say plainly that no scan is ever a guarantee. Anyone who knows what a scanner looks for can write around it, which is why I read the code as well rather than just running a scan.

Nearly always because the visible files were cleaned while the way in stayed open. The same vulnerable plugin is still installed, or one backdoor file survives in a directory nobody inspects. Given time, they use the same route again. That is exactly why the middle package is built around finding the entry point rather than around cleaning.

It depends on the case, and I will tell you plainly if the evidence points to data actually being pulled out. That matters legally: if personal data was exposed you have notification duties, and staying quiet because you are afraid of losing customers is a much larger risk than the breach itself.

Yes. I set the name in type so the site looks finished without a logo. If you want a proper logo, hire a designer for that specific job. I do not take logo work, it is not what I am good at.

Either way works. If I handle it, it goes in your name, never mine. You can move it to another host or another developer at any time without asking me.

The CMS package has an admin area for that from day one. On the other two, send me the new text and I change it free during the 60 days after launch. After that it falls under the monthly care plan.

No, and be careful with anyone who says it will. What I do is build the structure correctly, give Google the information it looks for, and keep the site fast. Position depends on your competitors and on time. Usually you start seeing movement in 1 to 3 months.

No, I use Next.js. For a site this size it is faster and there is far less to keep patched. If you specifically need WordPress, hire someone who works in it every day. You will get a better result than you would from me.

One developer in Thailand. No team, no agency, no registered company behind it. The contract states the hours I answer in, converted to your time zone, so you know when to expect a reply. Anything sent outside those hours waits until the next working day.

You get an API with documentation you can fire test calls from in a browser, so your team works without asking me one question at a time. We agree the shape of the data in one round before anything is written, so neither side has to rework later.

Those are finished products that arrive with a front end, for owners who want something they can switch on. This is the back end only, for people who already have a front end or a team building one. If you have nothing yet, I will point you at those first because they finish sooner and cost less.

No, and I sign an agreement saying so. The work ships under your name and the documentation is written so you can hand it straight to your client, with my name nowhere in it if you prefer.

TypeScript and Node primarily. PHP with Laravel and Python are both fine, and if your existing system is one of those I continue in it rather than making you switch to what suits me. Sample code in all three is public and readable before you commit.

Judge it on three things you can check: tests that run and pass in front of you, documentation somebody else can work from, and the security review included in Harden. Security work is my main line, so I review my own work before it ships.

A spreadsheet works right up until a second person edits the same file, or one product has two units. Buy a box of 24 and sell singles, and the file is left holding a number nobody is sure about, boxes or pieces. It goes wrong with nothing to warn you. The system converts units on its own and records who touched what and when. Those are the two things a file cannot do.

Yes. Tens of thousands is still normal. The part that needs care is search, not the count. You can find things by name, by code, or by scanning, and I load your existing list in before handover.

The staff screen has three buttons: receive, take out, and find. It is built to scan with a phone and be done. Reports and cost prices live on the owner side, which staff never see and never have to learn.

Yes. What each package includes is written out on this page, and there is a plain list of what is not included. Anything outside that list gets quoted before it happens, and you decide. Nothing is built first and billed afterwards.

Yes, immediately, if the sale goes through this system, from Count upward. If you use somebody else's till software, it depends on whether that software allows outside access. I check at no charge and tell you before we start whether it is possible, rather than promising and discovering later.

Control works out cost on a moving average, which answers which items earn and which ones to drop, and that is enough to run the shop on. If your accountant needs FIFO with layered batches, that is a different size of job and it sits in Custom. I say so here rather than letting you find out at year end.

Yes. The code is yours, inside your own account from day one, with documentation another developer can pick up. I do not lock clients in by hiding how it works.

If your work does not lean on the camera and does not have to run without a connection, I will tell you to build a website. It costs less and it is easier to change later. I say that during scoping. I do not push an app because an app bills more.

No. Building for iOS needs a Mac and I do not have one. That is a tool limit, not a skill one, and I am saying it here so it does not come up at the end of a conversation. If you need both platforms, hire someone separately for iOS.

Yours. Open the Google Play developer account in your company name from the start. Then the app is your property and you can change who maintains it whenever you want, including replacing me.

In Scan, images are read on the phone and no picture leaves the device, which suits confidential paperwork. In the other two packages, photos go to your own server over an encrypted connection. Never to mine.

ReceiptLens. A native Android app I wrote in Kotlin that reads receipts with the camera and does the processing on the phone. The code is public on GitHub and it is on the portfolio page.

Four things: how many screens, whether it has to keep working with no signal, how much work the camera does, and whether you need a web console. A five-screen app reading from your server, and an app that runs offline, reads documents on the phone and feeds a console, are not close in size. The three prices above are fixed, and the scope behind each one is written down before anything starts, so nothing gets added on later.

Because I pay a full year of domain and hosting up front in the first month. On a month-to-month deal that stops halfway, that cost lands on me. So I set the term at the start and write it into the contract, rather than mentioning it once you are already in.

Yes, after one review pass. I need to see whether it can be maintained as it stands and how much technical debt is sitting in it. If the honest answer is that it should be rebuilt rather than maintained, I will say so, even though that is a smaller sale for me.

You do, all of it. The domain is registered in your name from day one, not mine. I export the data and hand over the source code. Nothing gets held back to keep you tied to me.

Per issue, not per message. If one issue takes ten messages to sort out, it still counts as one. I do not want you hesitating before writing to me.

Monitoring runs 24 hours and alerts me, and I work late as a matter of habit, so I usually know before you do. But I do not guarantee a response outside business hours. I would rather not promise something I cannot hold to every time.

I am at UTC+7, and the reply window in each package is counted against your business hours, not mine, which is written into the contract. In practice you often send something at the end of your day and find it handled at the start of the next one, but the promise is only the stated window.

No. By default I test in a way that does not disturb live traffic. If there is a check that could affect the system, I ask permission first and schedule it for your quietest hour.

Not for the external review. For the authenticated test I ask you to create a fresh test account instead of sharing a real one, and you can delete it the moment I am finished.

Yes. It is written in two parts. The first is for you: what the risk is and what to deal with first. The second is for whoever does the fixing, with the technical detail. From the authenticated test up there is also a video where I talk you through it.

Yes, and the report is written so they can. It is all in English, and I do not hold anything back to keep you tied to me. If you would rather I did the work, that is the third package.

No. I am one developer working alone, with no company behind me and no security certificate on the wall. What I can show you instead is the code. The tools I test with, raidkit and raidscan, are public on GitHub, and I will send you a sample report before you pay anything.

I am in Thailand, GMT+7. I answer within one business day and the reply window goes in the contract converted to your time zone, so there is nothing to guess at. Everything runs in writing, and no call is required at any point.

A number given before anyone knows the real scope is always wrong, and you pay for it either way: through overruns, or through me quietly cutting quality to stay inside the budget. So I quote one stage at a time, and the number for that stage is fixed and in writing before it starts.

For work this size I bring in people I have worked with before, but I am the only one responsible and the only one you talk to. Staging the work is part of the same answer: you are not betting the whole project on me on day one.

I work alongside them, not over them. The code goes into your company's own repository accounts, and the README, comments and documentation are in English so your developers can pick it up and keep building. The goal is that they don't need me.

You stop. You still have one working piece of software and all of the source code, so someone else can carry on from there, and the 60 days of support still runs from the day it went live. That is the trade against signing for everything at once and waiting a year to find out.

Sign an NDA before any details change hands, that's normal and I have no problem with it. The design also keeps your data in your company's own accounts from the start, not on anything of mine.

I'm a developer working on my own from Thailand. No company, no team, no sales staff, and no security certifications, so if any of that is a requirement on your side it's better to know now. Before anything starts I give you the hours I'm reachable, converted to your time zone, and I keep to them.