Rubytech LLC
We build software for organisations with complex, fragmented data. Our work improves process efficiency, reduces operational risk, and improves reconciliation.
Most engagements begin with a question about how a business operates and end with software in daily use.
We document how the business operates today, where the time and the cost sit, and what is worth changing. Where the technology is not the cause of the problem, we say so.
We design the user interface and build the systems behind it. Designers and engineers work on the same engagement.
We start products of our own and operate them. Two of them run day-to-day operations in construction and estate agency.
Available singly or together. Where an engagement ends with a system in operation, Rubytech will either operate it or train the client's own staff to.
We document the current process, specify where it should change, and sequence the changes with costs against them. The changes are ranked by value against effort, and the plan is written so that another party could implement it.
Outcome: a costed, sequenced planWe design the product and build it: the user interface, its behaviour, and the systems behind it. Designers and engineers work on the same engagement. A prototype is tested with users, and the results revise the design before the build.
Outcome: a product in daily useWe consolidate records held in separate systems into a single store, so that one customer is one record and every system reads the same facts. Where the records disagree, that is documented first. The three services below depend on it.
Outcome: a single source of truthWe apply AI to repetitive process steps: chasing, re-keying, and checking one list against another. It runs under human review and keeps a record of what it did. Where AI does not pay for itself, we say so.
Outcome: fewer steps done by handThrough RubySDK, our Web3 division, we design and build private markets for assets that do not have one: property, art, watches, wine, coins. A token divides the asset into parts, and a transfer settles directly between the two parties.
Outcome: an asset traded in partsOwnership records and transfers can be built so that only the parties can read them, and so that a transfer to someone not on the approved list does not settle. An accredited party reviews the contracts before anything goes live.
Outcome: privacy by cryptographyRubySDK, the Web3 division
Most tokenisation to date has covered financial instruments: funds, credit, treasuries, equity. RubySDK works on the assets outside that set: a building, a painting, a case of wine, a watch, a coin. Assets held outside the financial system, owned outright, and ordinarily bought and sold whole.
A token divides the asset into parts, so an owner can sell a quarter of a building and retain the rest. Each part is held in its owner's own wallet, with no custodian and no platform between the owner and the ownership record. A transfer settles between the two parties.
The client owns the marketplace the end user sees: the listings, the pages, and the buying and selling process, built from components Rubytech supplies. Rubytech owns what sits underneath: the token, the record of who is permitted to hold it, the settlement of a transfer, and the decentralised privacy infrastructure it runs on.
Some networks conceal balances and transfers as a property of their cryptography. Some are transparent and are wrapped in a cryptographic privacy layer. Some are permissioned and separate the data, so that only the parties to a transaction hold it on their own machines. Which of those suits an asset is settled in an architecture phase, along with the token standard, the chain and the contract language.
The same stack carries systems other than ownership registers. Rubytech has built complete ecommerce systems on it, in food and drink and in apparel.
All four can be driven programmatically or through a user interface, so your own front end controls them.
Fixed-price engagement
A fixed price for a fixed scope of work, with a start date and an end date. It sets out what is worth building, what is worth buying, and what should be left as it is. There is no open-ended discovery phase and no retainer.
We document how the work is currently done, identify where time and cost are lost between systems, and rank the options by value against effort. The output is a written report and a costed shortlist.
The technology is selected against the requirement rather than decided in advance. Where an engagement should not result in a build, the assessment says so.
How the process is currently run, who the parties are, and where time and cost are lost. Technology selection follows this stage.
The changes are ranked by value against effort, and the ones that would not pay for themselves are recorded as such.
Where the data is held, and, where ownership has to be provable and holdings private, which network and which token standard. Tested against the requirement before anything is committed.
A working version is built in a sandbox and tested with users. The results revise the design before the finished build.
Everything in production has a named owner, a documented way to switch it off, and a review cycle after go-live.
All four are live and built principally on Anthropic's Claude models.
The AI environment installed on a client's own server: Claude Code joined to a persistent knowledge graph holding every person, document, conversation and decision the client has touched, with email, messaging, research and document production connected to it.
Claude Code · knowledge graph · agent orchestration getmaxy.com → EducationA course, in London and online, that teaches knowledge workers to set Claude up around their own work without writing code. Twenty-eight modules across five stages, from the basics of the models, through bringing archives into a structured data layer, to running the setup day to day.
Claude fundamentals · data layer · applied practice maxy.institute → Deployment · ConstructionA back office for small builders and trades, built on the same environment. It covers several concurrent jobs, and the quotes and approvals that move between parties who each hold their own version of the record. It runs under human review.
Small builders and trades · live deployment sitedesk.online → Deployment · Estate AgencyBuilt for UK estate agents, whose records are held across systems that do not connect to each other. It handles enquiries, listings, correspondence and the paperwork that progresses a sale. Same architecture as Maxy, configured for agency operations.
UK estate agents · live deployment realagent.network →No. Article 50 of the EU AI Act became applicable on 2 August 2026, and the claim that all AI-generated content must be labelled or the fine is €15 million is a significant overstatement.
Four duties are in force. Two are likely to touch you: tell people when they are dealing with a chatbot, and disclose AI-generated text published to inform the public on a matter of public interest.
The exemption most of the posts leave out covers the rest. Where a named person substantively reviews the content and takes editorial responsibility for it, there is generally no disclosure duty.
Read the view →Usage numbers tell you how much AI is being used but they do not tell you if anyone is using it well and improving.
The Compliance API carries the events and the conversation text, but not the tools that were called or the reasoning behind them. You get the final outputs but not the working.
The working is in the session's own log, which you can collect openly and with staff informed. What it cannot show is what they did with the output, so a complete assessment takes three lenses: the outcome against the AI's last draft, the session trace, and the person's own reflection.
Read the view →With the work, not the technology. The most impressive projects deliver the least, because they start with the clever thing and go looking for somewhere to put it.
Map a week of the job first, step by step, and the opportunities show up on their own: the stalls, the same details keyed in twice, skilled people on work that does not need them.
Then build only where an agent genuinely helps. Sometimes it does not, and going live is the start of the loop rather than the end of the job.
Read the view →Rubytech LLC is a systems engineering, data analysis and technology consultancy, organised as a consultancy, a design studio and a venture builder. Most engagements draw on more than one of the three.
The firm's people have worked in banking, insurance, healthcare, data and science, real estate, logistics and industrials, all sectors where data volumes and system counts are high. The work that carries across them is making a set of disconnected systems operate dependably, at any size of organisation.
RubySDK is the Web3 division. It designs and builds private markets for real-world assets: the token, the register of who is permitted to hold it, the settlement of a transfer, and the decentralised infrastructure it runs on. In that division the access control is enforced by cryptography rather than by procedure.
Rubytech's own products run on Anthropic's Claude models, and two of them run operating businesses in construction and estate agency.
Services are provided in the United Kingdom and across the EEA.
Contact us for further information with the nature of your request.