Bruncsoft — a small student studio that builds what you need Bruncsoft Home Services Why us Pricing Team Portfolio Contact A small student studio Cheap, fast, and honest about it. We build websites, front ends, back ends, e-shops, automation and AI tools — a static site is €80 and a dynamic one €200 , whole site, price agreed before we start. We also research and check source libraries for education platforms and companies that need to know where a claim came from. What we build See prices Get in touch What Bruncsoft is One thing, done properly There used to be twenty-six products. Most were half-built, or competing with companies that will always outbuild us, so they came down. What is left is the work itself — done for you, and finished. Built for you Websites, applications, e-shops, automation, AI tools and deploying AI products — plus hosting and maintenance so it does not rot. You talk to the person who writes the code. If we are the wrong people for it, we say so. All eleven things we do → Sources, found and checked For education platforms and for companies that have to know where a claim came from: we find the sources, index them so a search returns them, and check each one — including whether it is legally usable . Your library stays yours. How it works → Older work — Prompt Scrub, the chess model, MicroMem VLA, Bruncsoft Connect — is on the Portfolio page , mostly as open source. Start here Tell us what you need built. You talk to the person who will do the work — no account managers, no lost context. We will tell you what it would take, what it would cost, and if we are the wrong people for it, we will say so. Get in touch See prices Why choose us Why Bruncsoft Six reasons that are true today — not slogans. Bruncsoft is a student project, and that is part of the answer to most of them. What we actually build is on the Services page . We run what we build This site, its server, its database and its deployment are ours — on the same free tiers and the same rate limits you would hit. So the advice you get is advice we have already paid for, in AI bills, migrations and things that broke at the wrong moment. Built from scratch, not assembled No templates, no page builders, no third-party platform underneath. We understand what we deliver line by line, which is why we can still fix it a year later. You talk to the person building it No account managers and no brief passed through three people — you get answers from whoever wrote the code. Bruncsoft is a student project, so it is small and honest about being small: fast decisions, and a plain no when something is out of reach. AI where it earns its place We deploy AI daily — and we will also tell you where it does not pay off. Sometimes a rule beats a model; sometimes a small local model beats a cloud call. The token bill is part of the design, not a surprise afterwards. Price and scope up front Everything agreed before we start: clear scope, clear price, no hidden items. And when something is outside what we do, we say so instead of learning it on your budget. You own everything After payment the work is entirely yours — code, product, IP. No vendor lock: we can host and maintain it, but you are never stuck with us. Ready to get started? Contact us or see our services . Get in touch Contact Us Have a project in mind, or just a question? We'd love to hear from you. Write to us directly You can also reach us directly by email. We typically respond within one business day. Name Email Subject Message Send message Who is behind this Team One person, so far. Saying that plainly seems better than implying a floor of engineers. Viktor Brunclík Founder & Lead Builder Bruncsoft is a student project. It was started to learn how software really gets built and to earn something doing it — in that order — and everything on this site was designed, decided and shipped here. Get in touch → Looking for people to build with Bruncsoft is a student project, and it would be a better one with a few more people on it — not as schoolwork, as something that ships. Interesting to you if you like making things that other people actually use: front end, design, machine learning, writing, or being the person who notices what is broken. There are no salaries and no promises. What there is: real code in production, real users, and your name on it. Get in touch What things cost Pricing Real numbers where we can give one, and an honest “depends” where we cannot. Cheap because we are a student studio, not because something is missing. Website Static site €80 The whole site, not per page. Runs on GitHub Pages, so hosting costs nothing and there is no server to go wrong. - Design and build, start to finish - Hosted free on GitHub Pages - Edit your own text and pictures — free - Maintenance not included — available separately Get one Website Dynamic site €200 The whole site. A database behind it, working forms, and somewhere to log in and change things. - Database, forms, admin - Edit your own content — free - Hosting available, or bring your own - Maintenance not included — available separately Get one Afterwards Hosting & maintenance Monthly A site nobody maintains is broken within a year. Offered on both kinds, and you can stop whenever you like. - Updates, backups, someone to ring - Changes on request - Hosting too, if you would rather not Ask for a quote Built for you Apps, e-shops, automation, AI Per project Too different from one another for a price list. Scope and price are agreed in writing before anything starts — no hourly surprises. - Web applications and e-shops - Automation, on your machines or ours - AI tools, and deploying AI products - LiteLLM deployed, maintained, hosted Tell us what you need For education platforms Sources, found and checked Per project We research and index a source library for your platform, check that each one holds up, and check it is legally usable . Your library is yours — separated in the database, invisible to us elsewhere. - Finding sources, not just checking them - Liveness, dates, archive, publisher, corroboration - Legality checked too — licence, access, and no shadow-library copies - An API is planned Talk to us Prices are in euros and exclude VAT. A static or dynamic site is the whole site, not a per-page rate — a bigger or more complicated one we will quote before starting. Everything else is scoped and priced in writing first, and no work begins until you have agreed to it. What €80 actually buys → What we have shipped Portfolio Real things that run, not mockups. Where the work is open source, the link goes to the code. ExpireScan Live AI-powered expiring-domain intelligence — finds and watches domains as they come free, so a name worth having does not quietly get taken by somebody else. expirescan.com ↗ Prompt Scrub Open source A trained model that finds names, passwords, API keys, card and ID numbers in text and redacts them entirely inside the browser , so the data never leaves the machine. Pure JavaScript, zero dependencies. GitHub ↗ Chess move prediction Open source A model that predicts the move a human will actually play, not the engine-best one. Measured on a month it never saw: 38.5% top-1 and 64.0% top-3, against 5.5% for a random legal move. GitHub ↗ MicroMem VLA Open source Many small robot-skill models instead of one large one: record a demonstration, train a model small enough to read, write down what it does, and let a language model pick between them. GitHub ↗ Bruncsoft Connect Open source A self-hosted integration hub. It runs on the customer’s own server, issues its own key for every connected site, and never talks to us — there is no central service to phone home to. GitHub ↗ This website Built from scratch — no template, no page builder, no platform underneath. Multi-page build, own Express server, own SQLite, own visitor counting with a real consent gate. github.com/Bruncsoft ↗ Client work is not listed here — we ask before showing anybody’s project. If you want references for something specific, ask us and we will put you in touch. © 2026 Bruncsoft™. All rights reserved. Bruncsoft™ is a brand, not a company. 𝕏 @Bruncsoft GitHub Terms of Service Privacy Policy Cookie Policy Built with Claude Code · Powered by Vite · Hosted by Bruncsoft ✕ Terms of Service Effective date: August 25, 2026 1. Who we are Bruncsoft™ is a brand, not a company. “Bruncsoft” is a trademark and trading name used by Viktor Brunclík; it is not an incorporated company, and nothing on this site should be read as implying otherwise. Where these Terms say “Bruncsoft”, they mean that work and the person behind it. 2. Services Bruncsoft is a student studio . We build websites, web applications, e-shops, automation, AI tools, and we deploy and maintain AI products including LiteLLM . We also offer hosting and maintenance for what we build. Everything bespoke is delivered project by project as agreed in writing. We also research and check source libraries for education platforms and other companies — see clause 7. We no longer run a product of our own ; everything here is work done for a customer. Things we have retired. The AI Platform (BruncGate, Loop, Empathy), Bruncsoft Truth as a separate product, Bruncsoft Automation as a subscription, Terms & Docs, Bruncpedia, the two e-shops and the live AI chat demo were withdrawn during 2026. No customer data was deleted; if you had anything with any of them, write to us. 3. Websites: what a price includes A static website is €80 and a dynamic website is €200. That is the whole site, not a rate per page. Prices exclude VAT. A site that is substantially larger or more complicated than the usual case is quoted before work starts, and we say so before you commit to anything. What is included: design and build; free content editing for you through Decap CMS; and, for a static site, hosting on GitHub Pages at no cost, on an account that is yours. What is not included: maintenance . It is offered separately and billed monthly, on both kinds of site, and you can stop at any time. Without it, the site keeps working but nothing is updated, backed up or fixed by us. 4. Scope and pricing of bespoke work All engagements beyond the fixed website prices are individual. Scope, deliverables, timeline and price are agreed in writing before any work begins. No work is performed, and no payment is due, until both sides have confirmed the terms. If something turns out to be outside what we do, we say so rather than learn it on your budget. 5. Delivery and ownership Upon full payment, you receive full ownership of everything produced for you — code, documents and intellectual property. Bruncsoft retains no right to reuse or redistribute client-specific deliverables. There is no vendor lock: we can host and maintain the work, but you are never stuck with us. Where we review or extend work built by somebody else, that review creates no warranty over the existing work and no liability for problems already present in it. 6. Acceptable Use Our services may only be used for lawful purposes. We refuse to build, host or support anything involving illegal activity, harassment, deception of the people it is aimed at, or the circumvention of somebody else’s access controls. 7. Finding and checking sources We research, index and check source libraries for education platforms and for companies that need to know where a claim came from. This includes checking that each source is legally usable — its licence, whether it is freely accessible, and a hard refusal on shadow libraries and paywall-bypass services. We check sources, not claims. A report says what is verifiable about the page carrying a statement: whether it is reachable, dated, archived, who publishes it and whether other public indexes cover the same topic. It does not decide whether anything is true , and a good source can still be wrong. Part of a report is produced by a trained model. That is AI, it is labelled as such on every line it produces, and it can be wrong. The other checks are measured or come from a list a person wrote, and are never labelled as AI. Your library is yours. It is stored separately and is scoped to your account in the database, not merely by policy. A check is a report, not a guarantee , and not legal advice about whether you may use a given source. 8. Retired: Bruncsoft Edu We used to run a public study shelf at /edu . It closed in August 2026 — a handful of sources on a page nobody has time to maintain is worse than not having the page. Nothing was deleted and nobody lost anything : there were no accounts and no paid plans. The research and checking behind it is still offered, as the bespoke work described in clause 7. 9. Deepfake defence and security work What we do not promise. No control we design can guarantee that fraud will not occur, and we say plainly that deepfake detection tooling is unreliable on its own . Our work reduces exposure; it does not remove it, and it is not insurance. We are not a law firm, an auditor or a certification body — what we produce is practical engineering and training. Simulated phishing or impersonation drills are only ever run with written consent and within an agreed scope. 10. Hosting, maintenance and third parties Hosting and maintenance are billed monthly and can be cancelled at any time. Where we put a site on GitHub Pages , or content editing on Decap CMS and Netlify , those run under their own terms on accounts that belong to you — we set them up, we do not control them, and their availability is not something we can guarantee. The same applies to LiteLLM : it is somebody else’s open-source project, it runs on your infrastructure with your provider keys, and your model providers bill you directly. We are not a party to your agreement with them. 11. Free tools and open source Some of what we publish costs nothing and is on GitHub under the MIT licence. It is provided “as is”, without warranty of any kind . Software you self-host runs on your infrastructure, under your control and your responsibility; we do not operate it and cannot see it. 12. Our own analytics We count page views ourselves rather than using Google Analytics or any other third party. Nothing is counted until you accept the banner , and what we keep is which page, which day, and which site linked here. We do not store your IP address, your browser, anything that identifies you, or your search terms, and nothing is shared or sold. Declining costs you nothing and hides no part of the site. 13. Limitation of Liability Bruncsoft's liability is limited to the amount paid for the specific service. We are not liable for indirect, incidental, or consequential damages, nor for damage arising from AI output that was relied on without review, nor for free tools and open-source software provided at no charge. Nothing here limits liability that cannot be limited by law. 14. Contact Questions: ✕ Privacy Policy Effective date: August 25, 2026 1. Data We Collect We collect personal data in the following cases: - Contact form — your name, email address, subject and message. - Newsletter — your email address, until you unsubscribe. - Platform accounts — your email address, used to sign you in by emailed link and to issue an API key. We store only a hash of the key, never the key itself. Accounts are created by us ; there is no sign-up form. - Bespoke work — whatever you send us in order to have something built. It is yours, we process it to do the job, and it goes back to you. - Source checking — the URL you ask us to check and the check we produced. If you supply a claim to search for, we keep the normalised keywords rather than your sentence, so a verdict can be reproduced. Our own visitor numbers. If you accept the banner, we count which page was opened, on which day, and which site linked here. We do not store your IP address, your browser, your search terms, or anything that identifies you — a page view is a counter, so there is nothing here to link back to you. If you decline, nothing is recorded at all. There is no third-party analytics on this site. We also use browser localStorage — see our Cookie Policy . 2. Legal Basis & Purpose Contact enquiries and project work — performance of a contract, or legitimate interest in answering you before there is one (Art. 6(1)(b) and (f) GDPR). Newsletter — your consent (Art. 6(1)(a)), withdrawable at any time. Accounts and the content you put into a product — performance of a contract (Art. 6(1)(b)). Source checking for a customer — processed on that customer’s instructions as their processor (Art. 28 GDPR). Visitor numbers — your consent (Art. 6(1)(a)), given through the banner and withdrawable by clearing the cookie. We do not use personal data for marketing without explicit consent, and never sell it to anyone. 3. Data Storage & Retention Data is stored on our own server in the EU (Germany). Contact submissions are retained for up to 5 years for legal purposes. Newsletter addresses are kept until you ask us to remove them. Project and account data is kept for the life of the account and deleted on request. Visitor counts are aggregate from the moment they are written — there is no per-visitor record to delete. A request to delete your data is itself kept as a record that the request was made and acted on. We keep the fact, not the data. 4. Who Else Processes It (subprocessors) We keep this list short on purpose. Today it is: - Resend — delivers our transactional email (sign-in links, notifications). It processes the address and the message we send. - Hetzner — hosts our server in Germany. Our contact form is delivered by our own mail server, not by a third party. Source checking uses no language model at all , so nothing we check is sent to an AI provider. Public indexes we query. When a source is checked, our server fetches that page and asks Wikipedia/Wikidata, Crossref, OpenAlex and the Internet Archive about it. Those requests carry the URL or the search keywords and our own identifying User-Agent — never you, and never your account. Wikipedia extracts we display are licensed CC BY-SA 4.0 and are credited as such. For bespoke work , any third party involved (a host, a model provider, a CMS) is one you choose and contract with directly. We set it up; we are not a party to your agreement with them. 5. Your Rights (GDPR) Under GDPR you have the right to: access your data, correct inaccuracies, request deletion, object to processing, withdraw consent, and request a portable copy. To exercise any of these rights, email us at info@bruncsoft.com . We will respond within 30 days. If you believe we are processing your data unlawfully, you may lodge a complaint with the Czech Data Protection Authority (ÚOOÚ) at uoou.cz . 6. Automated Decisions We make no automated decisions with legal effect about you. AI in our products summarises, sorts and drafts; a person decides. 7. Data Controller Bruncsoft — ✕ Cookie Policy Effective date: August 25, 2026 1. What are cookies? Cookies are small text files placed in your browser by a website. They are necessary for certain technical functions and cannot carry viruses or install software. 2. Cookies we use We set no third-party cookies at all — no Google, no advertising network, no tracker. There are three of our own, and here is every one of them: - bs_analytics — our own visitor counting. Needs your consent. Holds 1 or 0 and nothing else. Until it says 1 , we count nothing whatsoever. Saying yes lasts a year, saying no lasts a day (so we can ask again rather than nag on every page). Declining hides no part of the site. - Session cookie — only if you sign in. A random identifier and nothing about you, httpOnly and secure , scoped to .bruncsoft.com . Strictly necessary for signing in, so it needs no consent. - Session cookie — if you rate something on Edu or vote on a blog post. The same cookie, holding a random key so one browser cannot vote twice. It carries no identity and we do not connect it to anything else. Browsing the site without signing in, rating anything or accepting the banner sets no cookie at all. 3. What we count, if you say yes Which page was opened, on which day, and which site linked here — as a running total, not a list of visits. We do not store your IP address, your browser or device, your search terms, or any identifier that follows you between pages. There is no per-visitor record, which means there is also nothing for us to hand over or lose. We do this ourselves rather than using an analytics service, because a service would mean another company processing your data and another name in our Privacy Policy. 4. Local browser storage We use your browser's localStorage (not a cookie) to remember your light/dark preference and which blog posts you liked or disliked. This stays in your own browser; we cannot read it from our side. 5. Legal basis Visitor counting — your consent (Art. 6(1)(a) GDPR), given through the banner and withdrawable at any time. Session cookies — strictly necessary for a service you asked for. localStorage — legitimate interest (Art. 6(1)(f)) in remembering your preferences. 6. Changing your mind Clear cookies for this site in your browser ( Settings → Privacy → Clear browsing data ) and the banner will ask you again. Clearing the session cookie simply signs you out. Withdrawing consent stops all counting immediately; what was already counted is an aggregate number with nothing in it that could be traced to you. 7. Contact Questions?