Tell usThe problem

Straight answers

What does DField Solutions do?

We build custom software end to end: AI automation, cybersecurity, blockchain, full-stack web and mobile, generative AI pipelines and web performance work. Most projects use two or three of those at once rather than one in isolation.

Who will I actually be working with?

Dezso Mezo, who runs the studio, and who is also the person who answers the phone. There is no account manager between you and the person building the thing.

Where are you based, and do you work remotely?

The studio is in Dunakeszi, just north of Budapest, Hungary. Clients are anywhere: the work is remote by default and we work in English, Hungarian and German.

How long does a project take?

The standard engagement is eight weeks. Weeks 1 to 2 are discovery, audit and quick wins already deployed. Weeks 3 to 5 build out the site and the AI workflows. Weeks 6 to 7 are security hardening and integrations. Week 8 is documentation, training and handover.

How does it start, and what does the first conversation cost?

A free 30-minute intro call. No obligation and no sales pressure. Describe the problem in plain words — you do not need a written specification to book it.

Do I own the code when it is finished?

Yes. At handover you get the code, the accounts and the documentation. Where we deploy tooling like n8n we self-host it on your infrastructure, so you own the automation too rather than renting it from us.

What does Purple Team mean?

Red team is the attack, blue team is the defence, and most agencies only sell one of them. We run the attack against your systems and then close what it found, so the hardening is measured against something real rather than a checklist.

Can you secure an AI agent against prompt injection?

That is a specific thing we do. If you are deploying agents that read customer input and can reach your data or your tools, we test whether they can be talked into leaking or acting, and then we close it. It covers the OWASP LLM risks, data exfiltration and auth bypass.

What do you build with?

PyTorch, LangChain, MCP, n8n, Claude and OpenAI APIs on the AI side. Solidity with Hardhat and Foundry on Ethereum, Polygon, Base and Arbitrum, and Anchor on Solana. Next.js and React on the front, Python, FastAPI, Node and PostgreSQL behind it, Docker and CI/CD around it.

Do you have your own products, or only client work?

Both. Four products are live in the market — UseAIEasily, JuriSafe, SimulSpeak and AI Chatbot Maker — all built end to end by the same studio that would build yours.

What if I do not have a specification yet?

That is the normal case. Tell us the problem in plain words on the intro call and scoping it is our job, not homework you have to finish before you are allowed to talk to us.

Can you take over an existing codebase?

Yes. Audits, performance rebuilds and security hardening on software somebody else wrote are a large part of the work. The first two weeks exist precisely to find out what is really there.

Can you do one piece of work instead of the full eight weeks?

Yes. Eight weeks is the shape of a full engagement, not a minimum order. A single security audit, one automation, a performance rebuild or a contract review are all normal pieces of work on their own. Say which one you want on the intro call and we will scope that rather than talk you into the whole plan.

What actually happens in the first two weeks?

Discovery and audit, and quick wins deployed rather than queued. That means we find out what is really running — not what the documentation says is running — and ship the fixes that do not need a decision: speed problems, the first automations, the obvious security gaps. You see working changes before you see a plan.

How will I know it is going well while it is being built?

You watch real progress instead of reading status reports about it. Work goes to a live environment as it is finished, so what you review is the thing rather than a description of the thing, and a week with nothing to show is visible immediately rather than at the end.

Do you work with agencies and other studios?

Yes, in both directions. We take on partner work for AI, security and blockchain, and we subcontract design and illustration. If you are a studio with a client who needs something outside your bench, that is a normal arrangement and not a competitive problem.

How do I know your security work is real and not a scanner report?

Because the deliverable is the exploit, with reproduction steps, and not a four-hundred-item export with the informational findings left in. Anything found gets fixed and then re-tested against the same technique, so what you receive is a verified control rather than a recommendation to consider.

We already have AI in production and it is misbehaving. Can you help with that?

That is a common starting point and usually a better one than a blank page. Something in production has real traffic, real failures and real logs, which is far more to work with than a specification. The first step is finding out whether the problem is the prompt, the retrieval, the model choice or the surrounding architecture — those look identical from the outside and have completely different fixes.

Which blockchains do you work on?

Solidity on Ethereum, Polygon, Base and Arbitrum, built with Hardhat and Foundry, and Anchor programs on Solana. That covers token contracts, DeFi integrations and Web3 frontends. If your chain is not on that list, ask — the answer may still be yes, but we would rather say so than find out halfway through.

Do you build mobile apps?

Yes, alongside web and backend work. Most of what we build ends up being used on a phone whether or not it is an app, which is why the web work is mobile-first by default rather than adapted afterwards.

Do you host and run things, or hand everything over?

Handover is the default and it is a real handover: the code, the accounts and the documentation, with tooling like n8n self-hosted on your own infrastructure so nothing depends on us continuing to exist. If you would rather we kept running it, that is a separate conversation and an honest one, not the outcome we engineer toward.

There is a free AI security self-audit on your site. What is it?

Fifteen questions, about five minutes, and nothing leaves your browser. It is a checklist rather than a scanner: it tells you which of the common gaps in an AI deployment you have not thought about yet. It is free, there is no sign-up, and it is genuinely useful before you talk to anyone — including us.

I am not sure which of your services I need.

That is normal and it is our job, not yours. Most projects use two or three of the six at once — a performance problem often turns out to be an architecture problem, and an automation project usually has a security question inside it. Describe the symptom on the call and we will tell you which bench it belongs on, including when the answer is that you do not need us.

Do you write about what you do?

Yes. The lab has write-ups of real work, and the glossary defines the terms this work involves without selling anything. Both exist so you can judge whether we know what we are talking about before you spend an hour on a call finding out.