From audit to long-term operation: software for data, documents and transactions.
We design and build web platforms, AI analytical systems for large document sets and integrations into enterprise infrastructure. Larger projects rest on written specifications, automated tests and operational monitoring.
Four areas we specialise in — from architecture and implementation to operation and further development.
●IT architecture & consulting
Technical audits, design of systems architecture and data models, choice of infrastructure and technology. Written specifications before implementation, process automation and integration of third-party tools. Long-term strategic collaboration as well as one-off consultations.
●Digital asset management & web portals
Development and long-term operation of industry portals with tens of thousands of profiles, as well as price lists and catalogues — data model, import flows, search, content management and SEO. Transactional platforms and marketplaces, including non-standard auction and pricing models and integration with payment systems. Valuation and brokerage of domain portfolios, and technical due diligence of software projects.
●Enterprise integration
Deploying software into a client’s existing infrastructure — on-premise, private cloud or hybrid. API design, integration with internal systems, access control, security architecture and data migration.
●AI analytical systems
Design and development of systems that combine deterministic data processing with language models: structured extraction, multi-layered analysis of large document sets and conclusions linked to specific sources. The output is working material — the decision stays with a person.
Our current main focus: processing and analysing large volumes of structured and unstructured data — from industry portals with tens of thousands of profiles to the ActaIntel analytics platform for large case files and data rooms.
Platform · technical development
ActaIntel: an analytical layer over case files, data rooms and regulation
Legal intelligence. Ad acta. · Operator: Ad Acta Advokátní kancelář, s.r.o. · Technical development: Mercury Project
ActaIntel is a private analytics platform for lawyers that turns large case files and data rooms into a structured working memory of the matter — always anchored to source. It brings together the Ad Acta law practice and our technical development: the firm shapes the legal layer, and we build the technical layer. The platform grows out of the firm’s own legal practice; it is not a laboratory demo.
Case working memory
One verifiable model of the matter for the whole team
Documents, actors, events, deadlines, evidence, contradictions and legal questions in a single structure. Partner, senior, associate and paralegal work from the same material, each with access by role.
Not only “what applies”, but “from when, to whom and with what consequence”
For legislative and regulatory work, ActaIntel does not rely on a general-purpose language model alone. It works with its own local library of official Czech legislation from e-Sbírka, the electronic Collection of Laws, including historical and future-effective versions — with version diffs, transitional provisions and linked European context.
A public library of legal principles from Czech Supreme Court and Constitutional Court case law sits on the same analytical layer as the work on client files. Every citation links to the court’s official database and shows its verification status.
Supreme CourtConstitutional Courtverification status
Technical properties
Anchored to source
Every material claim is linked to a specific document and page.
Audit trail
At document, task and user level, including output versions and review steps.
Access control
Role-based; every document request is authorised against the specific matter.
Document integrity
Every uploaded file carries a SHA-256 checksum to verify its integrity for as long as it is stored.
Processing
Separate asynchronous queues for documents, analyses and audio transcripts, with automatic retries on transient failures.
Storage and models
Private EU storage; client data is not used to train models.
ActaIntel prepares analytical material and working drafts. The lawyer using the platform remains responsible for the professional assessment of its outputs and for the legal services they provide. Access is by personal invitation.
Apps for iPhone, iPad and Mac. Native Swift, no third-party libraries.
We also build our own apps for iPhone, iPad and Mac, from design through release to maintenance. They rely solely on Apple’s frameworks (SwiftUI, HealthKit, WidgetKit, RealityKit, SwiftData, CloudKit, IOKit), with no ads, analytics or user accounts. Data stays with the user — on the device, or in their own iCloud.
No subscriptions
No ads
No tracking
No account needed
Data stays with the user
MoveStory
Steps & activity · iPhone
The story of how you move. Steps and distance first, every activity counted, and your whole history at a glance.
Steps and distance in kilometres and miles, today’s steps right on the app icon, plus Home Screen, Lock Screen and StandBy widgets.
Every activity counts — walking, running, cycling, swimming, skating, rowing and more, from Apple Health.
A meadow that blooms as your day fills up, and a goal that follows your own last 30 days. No streaks, no guilt.
Free · No ads · No tracking · Health data never leaves your iPhone
A calm menu bar monitor for your Mac. Quiet when everything is fine — and it tells you when something isn’t.
CPU, GPU and memory in the menu bar — colour only ever means something: green is fine, orange needs a look, red needs attention.
Knows when your Mac works hard — sustained heavy load turns it orange, and red if it keeps going on battery. It never guesses why: Activity Monitor is one click away.
Core by core, GPU by app — cores grouped the way your chip is built, the apps keeping your GPU busy, and memory pressure instead of a misleading “free RAM” figure.
One-time purchase · No network access · Apple silicon Mac, macOS 26 or later
Most of our projects are not named publicly, so instead of a list of references we show the kinds of work we have long been doing.
Industry portal
A platform with a large catalogue and long-term operation
Design and development of the portal, its data model, import flows, admin, search, content and SEO over data in the tens of thousands of records.
tens of thousands of profilescontent + SEOlong-term operation
Transactional platform
Auctions and marketplaces with automated payment handling
Auction and pricing logic, automatic payment matching via bank APIs, card payments with refunds and signature-verified payment notifications, QR payments, and invoices and other documents as PDFs.
payment matchingcard and QR paymentsPDF documents
Integration
Deployment into a client’s existing infrastructure
Architecture, APIs, security design, data migration and connection to internal systems — on-premise, private cloud or hybrid.
API designdata migrationhybrid environments
Scope of engagement
We take on small technical problems as readily as large projects, with responsibility for operation and further development.
Quick intervention
Audit, fix, second opinion
A short brief that needs a fast technical assessment, a fix, a proposed approach or due diligence before the next decision.
Project
A new tool, integration or portal
A standalone implementation, or taking over part of a system: architecture and a written specification, development with automated tests and a reliable deployment.
Long-term operation
Platform development and data work
Ongoing collaboration on operation and development: availability monitoring, dependency updates, data flows, new features, digital asset management and the technical direction of the product.
Process and standards
How we build and run software
We work to our own written engineering standards for architecture, security, performance, accessibility and data handling. They apply to our own products and to client systems alike, and we keep extending them with lessons from operations.
●Specification and review
Larger changes start with a written specification that is reviewed before implementation; the finished code then goes through a separate review. A change counts as done only once it has been verified on the running system.
●Tests and releases
We maintain thousands of automated tests across our projects; for key projects, continuous integration runs them on every change. The Mercury Labs browser games are released reproducibly, and the previous version can be restored with a single command.
●Operations and monitoring
Automated monitoring checks the availability of our key production sites and services and the validity of their certificates, and raises an alert on an outage; server load is tracked continuously. Applications and their background workers run under unprivileged accounts.
●Dependencies and supply chain
Third-party libraries in the applications we run are checked automatically every day against known-vulnerability databases; an incomplete check is reported as an error, never as a clean result. Routine updates are adopted only several days after release, which reduces the risk of supply-chain attacks; critical security fixes take priority.
About us
Mercury Project since 2009
Mercury Project has been active in the Czech market since 2009, building on its founder’s IT practice going back . We specialise in developing and operating web portals, managing digital assets and, in recent years, designing and building AI analytical tools for processing complex data. We also build and publish our own iPhone apps.
Over the years we have delivered projects for private companies, banks, government bodies, law firms, healthcare providers and academic institutions in the Czech Republic and abroad. We operate industry portals with tens of thousands of profiles and manage digital assets for clients across sectors.
Our analytical workflow has also been piloted in an academic setting, in collaboration with researchers from Brown University and Charles University.
Need a quick assessment of a plan, a fix for a problem, or a whole system designed?
We can start with an initial consultation, a technical audit or a small pilot, and follow up with a full project and long-term collaboration.
Mercury Labs is our playground. We build games here, try out new technology and hide surprises. Both games run right in your browser — free, with no sign-up.
For kids aged 3–7
Mercury Kids
A space learning game: numbers, letters, shapes, logic and listening — playful and calm. Trace real constellations, explore the Star Atlas and set up profiles for several children. An iPad and iPhone app is on the way.
Eight planets are hidden on this page, each with a small surprise. The hints reveal the secret shortcut — or just launch a planet with its button. How many can you find?
0 / 8 planets
MercuryHidden
Hidden orbit
An old-school cheat code: ↑↑↓↓←→
VenusHidden
Liquid metal
Rest your mouse for a moment — mercury pools around the cursor. Then move it.
EarthHidden
Back to 1998
The past is clickable: find “the late 1990s” in the About section.
MarsHidden
Mainframe
Triple-click the big star at the top (on a phone, press and hold it).
JupiterHidden
Warp speed
Type warp anywhere on the page.
SaturnHidden
Shooting star
Every now and then one streaks across the sky. Catch it!
UranusHidden
Barrel roll
Type roll. Mercury Squadron pilots know.
NeptuneHidden
Constellation
Type star.
Pluto · bonus
???Locked bonus
Find all eight planets to unlock a bonus.
Dwarf bonus
Pluto is still a planet around here. Launch the celebration!