Cube turns the way you already think about markets into an AI agent that runs for thousands of subscribers, in their accounts, with your perspective.
Every day, a thousand smart writers publish a thousand smart calls. Their readers nod, screenshot, share. And then they go back to passive index funds.
Executing the thesis is hard because sizing is personal
The writing you already publish: your reasoning, your rules, your market memory.
An LLM-native agent encoded with your method, voice, citations, and risk frame.
The agent runs for thousands at once. They keep control. You keep the strategy.
Substack
Paid and private content stays first-class. Cube ingests subscriber-only essays, free previews, comments, and archive metadata with the original page attached, so gated analysis can shape the agent without becoming loose scraped text.
YouTube
Long-form video becomes searchable source material. Cube pulls videos in with transcripts, chapters, timestamps, and channel context, turning spoken research into cited memory the agent can reference back to the exact moment.
Twitter
Public posts capture the real-time edge. Cube organizes threads, replies, quoted posts, and market reactions into time-aware context, separating durable beliefs from the fast commentary that shaped them.
After your sources are ingested, Cube creates the subscription layer around them: a branded product page, subscriber access, billing, usage limits, and payouts.
Cube turns published judgment into action only inside clear constraints: custody, risk, auditability, compliance, tax handling, and platform guarantees.
Risk caps are enforced by deterministic policy checks, separate from the agent's reasoning.
Cube has no withdrawal authority; only OAuth-scoped trading. Subscribers' money lives in their broker, full stop.
You, the subscriber, and Cube can each pause execution instantly. Halts are auditable and per-account.
OAuth status, permissions, stale tokens, broker outages, and rejected orders are watched continuously. If the connection is unhealthy, execution stops.
Before any trade, Cube checks cash, margin, holdings, market hours, instrument support, and account-level restrictions. No impossible orders leave the system.
Immutable per-trade ledger: what, why, which letter triggered it. Subscribers see it. Regulators can see it.
Lot selection, wash-sale avoidance, and year-end harvest opportunities are calculated per subscriber, not as a one-size-fits-all batch.
Every reasoning step the agent took, including the prompts, model versions, and data inputs, is queryable by you, your subscribers, and counsel.
We onboard creators one at a time, by hand, so every agent launches with the right archive, rails, pricing, and subscriber experience.
No.
Cube is not designed around manual trade approval. Creators do not review every subscriber action before it happens.
Instead, you define the product up front: the source material, strategy logic, constraints, eligible actions, risk boundaries, and subscriber experience. Cube operates inside that approved framework.
Your role is to own the thesis, the process, and the ongoing published view. Cube handles the translation from that view into a structured product subscribers can follow.
Markets move. Losses will happen.
Cube does not promise outcomes, and subscribers are not buying a guarantee. They are subscribing to your investment process, your research, and your way of thinking through markets.
Cube helps make that process legible. Actions are tied back to your published thesis, stated rules, portfolio logic, and relevant market context, so subscribers can understand why something happened, especially when performance is difficult.
The goal is not to avoid every loss. The goal is to make the reasoning clear enough that trust can survive volatility.
Cube is built around a creator-defined framework.
Before launch, we work with you to define what the product should and should not do: which sources matter, what kinds of actions are allowed, how risk should be handled, what constraints apply, and how your views should be interpreted.
If your thinking changes materially, the product framework can change with you. Cube is not meant to replace your judgment. It is meant to make your judgment usable, consistent, and clear to subscribers.
Copy-trading centers the creator’s trades.
Cube centers the creator’s reasoning.
A copy-trading product asks, “What did this person buy or sell?” Cube asks, “What does this creator’s current view imply for a subscriber, given the framework, risk limits, and context?”
That matters because your product is not just the trades. It is the thesis, the timing, the conviction, the risks, the updates, and the moments when the right move is to wait.
Cube gives creators a way to monetize the full investment process, not just a feed of transactions.
You can. Trade alerts are simple.
But they leave a lot of value uncaptured.
Subscribers still have to answer the hardest questions themselves: how much to size, whether it fits their portfolio, what changed since the original call, whether the thesis still holds, and what to do next.
Cube turns your ongoing work into a more complete subscription product. Instead of selling alerts, you are selling access to a structured version of your investment process.
No.
A traditional algorithm follows a fixed rulebook. Cube is built around your evolving body of work: essays, updates, models, memos, watchlists, portfolio notes, and stated constraints.
Your archive helps establish the foundation. Your ongoing writing keeps the product current.
The point is not to flatten your judgment into a rigid system. The point is to turn your market thinking into a product subscribers can follow over time.
No.
Cube handles the technical work.
We ingest your sources, structure the product, configure the subscriber experience, set up billing and access, and support launch. You bring the research, judgment, and audience. Cube handles the product layer around it.
Onboarding is designed to be fast.
During private beta, Cube works directly with each creator to ingest sources, define the product, configure controls, set up the subscription page, and prepare for launch.
For creators with a clear archive and focused strategy, setup can move quickly. The exact timing depends on the complexity of the product and the source material.
Payments are handled through Stripe, similar to Substack.
Cube helps set up the subscription product, billing, and payout flow during onboarding. Subscribers pay through Stripe, and creator payouts are managed through Stripe.
Yes.
Creators set their own pricing, with guidance from Cube during onboarding.
We can help think through packaging, tiers, positioning, and what subscribers are likely to expect, but the creator owns the commercial relationship.
Not during private beta.
The private beta is focused on paid subscription products for selected creators. Free tiers and trials may be introduced later, but they are not part of the initial launch.
Cube handles the product infrastructure around your investment thinking.
That includes source ingestion, product setup, subscriber access, billing through Stripe, usage controls, permissions, payouts, and the subscriber-facing experience.
You do not need to build software, manage payments, or stitch together tools. Cube gives you the layer between your published work and the subscriber product.