Your cloud
One platform.
Your own cloud inside it.
The same system serves a first hypothesis and a fully dedicated enterprise deployment. What changes as you move up is not the product — it's how much of the cloud underneath it is yours alone.
The architecture
Built as planes, not a pile.
Start at the plan you hold and follow it down: the infrastructure it runs on, the compute that scales per client, and the one data core — fed by live market-data integrations — that all of it pulls from.
Shared plans (Hypothesis · Experiment · Thesis), Enterprise plans (Teams · Enterprise), and the Enterprise White Label plan — your brand, your domain.
Shared cloud with a dedicated database from Experiment up; a dedicated, isolated cloud for Enterprise; the identical dedicated cloud for White Label, plus branding, domain, and custom-built features.
Shared or dedicated, compute grows the moment you ask for more — as fast as you want to go.
Ingestion, cross-checking, the indicator engine, and delivery — one engine between the feeds and every client cloud.
Ten independent feeds, cross-checked against each other: equities, options and Greeks, crypto venues, fundamentals, macro and rates, filings and insiders, short interest, dark-pool activity, corporate actions, and news.
Connected under your own keys, straight into your workspace — never into the shared core.
Two clouds, one product
Shared where it helps.
Dedicated where it matters.
Shared cloud
Everything a community needs.
The communities, published theses, leaderboards, and the knowledge base live on the shared plane — the parts of the platform that are better because everyone is in the same room.
- Per-member data scoping enforced in the data layer
- The same data core and grading engine as every plane
- Where the Hypothesis tier runs
Dedicated cloud
A boundary that is yours.
From the Experiment tier up, your work moves behind your own boundary: a dedicated database first, then — for Enterprise and White Label organizations — a dedicated cloud of your own, end to end.
- Your own database — your research never shares a store
- Dedicated compute for each client, scaled to the pace you set
- Data walls you dial per unit inside your own org
The data core
Heavy once, fast for everyone.
Market data, options chains, fundamentals, and research knowledge are ingested, cleaned, and maintained once — in a core every plane reads. Your backtests and Protocols draw on the same curated history whether you're on the shared cloud or your own deployment.
What's yours stays out of it: feeds and subscriptions you bring connect under your own keys and live in your workspace, behind your boundary — the core never sees them.
- One curated history — every plane reads the same maintained market record, so results are comparable everywhere.
- Your feeds stay yours — bring-your-own data connects under your keys and never enters the shared core.
- One grading engine — paper and live, member and org, everything is graded by the same engine against the same record.
Performance
As fast as you go.
Backtesting and study cadence run on your dial — from tick-by-tick all the way out to daily, with one-minute resolution as the default — and the platform provisions the compute to match. A desk that reacts tick by tick gets the horsepower for it; a desk that studies daily doesn't think about infrastructure at all.
Cadence is provisioned per member and per organization — your pace never queues behind anyone else's.
From ticks to daily — you set the pace
Every rung is the same engine and the same data — only the resolution and the provisioned compute change.
What runs on it
The infrastructure serves the method.
Protocols
Agent-built indicators, screens, scoring systems, and studies — written on demand, sandboxed, and backtested before they can influence anything.
Streaming signals
Live market data flows through your signals and Protocols as the market moves — alerts and the thesis workspace update in real time.
The thesis workspace
Falsifiable claims with immutable evidence, validated forward and graded — the record that everything else exists to serve.
Screens & scanners
Scan live chains and markets for the setup you just defined — as a built-in screen or a Protocol the agent writes for you.
Bring your own data
Your feeds, subscriptions, and APIs connect under your own keys and stay in your workspace — never in the shared core.
Bring your own AI
Plug in your own AI vendor keys. Your models, your bill, your workspace — the platform orchestrates them.
Memberships
Isolation that scales with the plan.
Every step up the ladder buys a harder boundary. The product is the same everywhere — what grows is how much of it runs for you alone.
- Hypothesis individual Shared cloud Communities, theses, and backtests on the shared plane — full multi-tenant isolation enforced in the data layer.
- Experiment individual Dedicated database Your data moves into its own store. Paper trading, Protocols, and studies run against a database that belongs to you alone.
- Thesis individual Dedicated database + live loop The full lifecycle on your own store: real-time signals, the thesis workspace, and real fills audited against their simulated twins.
- Teams organization Isolated workspace Org data lives on its own plane — shared research inside the desk, hard separation from everyone else.
- Enterprise organization Dedicated cloud Your own isolated cloud, end to end — not shared compute — with data walls you dial per unit and isolation guaranteed in writing.
- Enterprise White Label organization Dedicated cloud + your brand The identical dedicated cloud Enterprise runs on — plus your branding, your domain, and features custom-built for you.
Plans and what each includes live on the Membership page — Enterprise and White Label conversations start on the Enterprise page.
Run it on your own terms.
From the shared cloud to a deployment that exists only for you — the founding cohort shapes what gets provisioned first.