Composable workstation
PlannedIndependent panels in a grid, saved layouts, a trade journal and notices.
An analysis workstation that you build from modules
Charting, tick data, historical testing and agents in one place. Six modules sit on top of the base, and each one is paid for separately.
The workstation
What it does, with no module
9 things the workstation does before you activate anything. The six modules add research: daily use does not depend on them.
Born on crypto, built for any market that produces a tick.
Charts and bar types
AvailableCandles and Heikin-Ashi; renko and range in preview at a fixed size. Tick, volume and composable layouts planned.
Tick, book and order-by-order data
AvailableTicks with the aggressor side, book snapshots, order-by-order events. Real time on crypto.
AI agents with approval
PlannedThey read the market and propose, reasoning in plain sight. You decide how much autonomy.
Historical tests and optimisation
FinishingTick-level, deterministic historical tests with walk-forward and a map of stable parameters.
The code you already have
PlannedPine Script, JavaScript, Python and R: all four produce the same object, with the same report card.
Market screener
PlannedComposable market screens: filters you save as scans, and a plain-language question instead of a query.
Market replay
PlannedA session plays back bar by bar, on the real data of that day.
Portfolio and journal
PlannedTrades and notes in one exportable record. It prepares the data, it does not compute tax.
The layer that runs through everything
Connect your own AI and query the ecosystem in plain language. The layer keeps growing: what it does today is not its boundary.
- Connect your own model, or use the local ones included: the choice stays yours.
- Ask in English or in Italian, and the ecosystem answers with objects you can redo.
- Every hypothesis from the model crosses deterministic rules before and after: none passes on its own.
- Every answer carries its own derivation path, and it stays in the record.
- It runs on your own data, inside your own workstation, and passes what it computes to no one.
Why it cannot diverge
There is one computation, compiled twice: for the server and for the browser. The chart you see and the test that runs cannot say two different things.
- The same window gives the same number, whether you are looking at it or redoing the count.
- The past is not recomputed: what was measured stays as it was.
- Every number carries the way to redo it: window, step, seed.
«The code that measures is the code that is sold.»
The engine, in one line
The technologies, and why those
statistics · the starting pointR
orchestration and analysisPython
critical pathsRust
workstation graphicsJavaScript · WebGL
R is the reference for statistics, Python holds the analysis together, Rust carries the tick. Each language sits where it is best.
One engine, three places where it runs.
The computation lives in one core, compiled for the server, for Python and for the browser. The code that measures is the code that is sold.
Rust core
Incremental transforms, causal resampling, flow aggregators and order matching at tick level. One bridge to Python, one to the browser.
nativeresearch and server
Python bridgePyO3
browserWebAssembly
Layered graphics
A WebGL engine of our own: trading inside the bar, volume profile, liquidity heat map, vertical book. The panel grid is composable.
renderingWebGL
workstationbrowser and desktop
Windows dependencynone
Data and replay
Ticks with the aggressor side, book snapshots, submission and execution events. The past replays deterministically.
tick seriesdedicated store
event simulationnanosecond
replaydeterministic
Why a workstation like this does not exist
Three things cost a lot and never show in a demo: that is why the base came first.
Order flow outside Windows
Tick-level order flow, in the browser and on Linux, is not the cut-down version of a Windows program.
Look-ahead bias, measured
No look-ahead by construction, and the evidence that there is none stays attached to the measurement.
Falsified models get published too
A model that does not hold up is published anyway, together with the reason it did not hold.
Every decision stays written down
An append-only archive: for each decision it keeps the hypotheses, the alternatives set aside and the derivation chain. You go back and see why.
- The hypotheses that were formed, including those that did not hold.
- The alternatives set aside, with the reason next to each one.
- The derivation chain, from the raw data to the conclusion.
Six modules, each with its own page
They are activated and paid for separately, and none of them is the centre of the platform. Each page says what a module does and what it does not.
The plans, item by item
Three bundles or item by item, on a yearly basis. Prices are set when the platform launches.
Build your own
The Workstation, plus only the add-ons you need.
Prices are set when the platform launches: people on the waiting list get them first.
coming soon- Workstationalways included · from launch day
Two modules
price being set
coming soonA fixed bundle, not a pick-and-choose: the Workstation with volumes — volume profile and footprint — and the order book with its heat map.
- Workstation
- Volumes
- Book and heatmap
from launch day
Notify me at launchEverything included
price being set
coming soonThe Workstation, every add-on, the copilot and the data.
- Workstation
- Volumes
- Book and heatmap
- Cycles
- Options
- The copilot and the data
from the fourth quarter of 2026
Notify me at launchEverything included + Algo
price being set
coming soonAdds advanced backtesting. Assisted coding, script import, the strategy library and algorithmic trading are declared in the roadmap as phases.
- All modules, the copilot and the data
- Advanced backtesting, available
- Assisted coding — roadmap phase
- Script import from other platforms — roadmap phase
- Strategy library — roadmap phase
- Algorithmic trading — roadmap phase
from the fourth quarter of 2026 · Advanced backtesting is available; the other four items are phases declared in the roadmap.
Notify me at launchEvery add-on is an optimised tool, built on the research we publish: reading the chart stays yours. With the copilot you can build any tool; redoing that research takes years.
What arrives when
The Workstation, Volume profile and Orderflow are there from launch day. Options come later, within the fourth quarter of 2026: we would rather say so here than let you find out after paying.
- Moving up or down a tier comes at launch, once payments are connected
- Fourteen days to change your mind: before launch the refund is full, after it you only pay for the days you used
- No AI usage meter
- Books are bought separately: they are not part of the subscription
Artificial intelligence: the keys are yours
- No usage meter and no markup on calls: there are no AI credits at all.
- Models of ours in the price: none. What you pay for is the place where you plug in your own.
- Your own Anthropic, OpenAI, DeepSeek or Gemini key: you pay your provider directly. The engine is in place; the screen to enter the key is a phase declared in the roadmap.
- On a self-hosted install you use your own local models, and no request leaves your network.
Demo on request
No free plan: we work with professionals. A demo is granted at our discretion, for a few weeks and with limits, to see the workstation live.
Your datafeed
Already have a data provider? Connecting it to the workstation with your own credentials is a phase declared in the roadmap.
Are you an institution or a fund? Let us talk.
Get in touchThe six lines, and their status
Six lines of work. Each row says what state it is in today, not what it costs: the price, where it exists, lives on the product page.
Research
Papers with method, data and replication code; abstract, DOI and BibTeX stay public.
in preparationPublications
Training
Recorded video paths, plus a live masterclass for hands-on work.
available, price not publishedCourses
Platform
The software that runs the seven-stage chain on historical series.
not released yet
Trading systems
Automated systems under study: the line exists, a catalogue does not yet.
in development
Publishing
The books and the executable notebooks that come with them, to re-run and verify.
availableThe books
Institutional
Commissioned research and work with institutions: the conversation is the first step.
activeGet in touch
Where it lives, and with which account
The platform lives at app.cryptoverso.net, and you get in with your Cryptoverso account: the same one you create here, with no second sign-up.
That address does not answer yet, so it is written here and not linked. Prices are set at launch: leave your email at the bottom of the page.
Statusprice not published
All in the browser, nothing to install.
The platform is web-based and runs in our cloud: open it, sign in and work. The desktop version comes later, with the same account.
In the browser, right away
Nothing to install and nothing to update: it opens from any computer with your account.
Your own space
Data, strategies and log live in your own space, and you export them whenever you want.
Desktop, in design
The desktop version follows the web one, with the same account and the same modules.
Mobile app, in design
The phone is for looking: watchlist, open positions and alerts. The analysis stays where the screen is big enough for it.
Be the first to know
The platform is not out yet, and prices are set at launch. Leave your email: we write when it opens, with the prices.
The notebooks of the preview chapters open without an account.