The Thesis
The desk is the product. The loop is the moat.
Why we're building an agentic trading desk for crypto — and what's actually running behind it today.
The problem is fragmentation
A serious crypto trader runs a stack, not a tool. Market data in one tab. Charts in another. A portfolio tracker that's read-only. A bot platform locked to one venue. Tax software that doesn't know your strategy. Each exchange's own UI for the actual fills. Seven subscriptions, seven logins, zero integration — and you are the integration layer, copy-pasting between tabs at the exact moment a spread is moving.
That fragmentation is not an accident. Every exchange is its own silo. Every tool owns one slice of the workflow and stops at its edge. Portfolio trackers can't execute. Bot platforms have no research. Research tools have no live book to act against. Nobody closes the loop — because closing it means doing all of it, well, in one place.
We are not competing with any one of those tools. We are competing with the fragmentation itself.
What a desk is
Every professional trading operation has a desk: the one place where research, execution, monitoring, and automation live together, ready to act. Not a tool you open — the surface you operate from.
Plutous, the multi-exchange trading platform, is that desk for crypto. One non-custodial surface across every venue you trade. You research on it, you execute on it, you watch your whole book on it, and you automate the parts that should run without you. The fragmented seven-tab stack collapses into one operating surface — and the cost it removes isn't subscription fees, it's context-switching at the worst possible moment.
Why now: the agent changes what a desk can be
In 2020, an “all-in-one trading platform” was a worse idea than the fragmented stack — it just meant seven cluttered tabs behind one login. What changed is the agent.
The desk doesn't only hold your tools now. It can operate them. An AI agent can read your positions, pull the funding spreads, size a hedge, and place it — through the same desk you operate by hand, on your own exchange accounts. The agent is the skill bridge: it lets you describe what you want and get a real, running strategy back, without a quant team behind you.
That is why we say agentic, and we mean it specifically — not “AI-powered” as a label, but the agent as a first-class operator of the desk. Today you run the desk; the agent interface that lets your agents run it alongside you is what we're building next. We won't describe it in the present tense until it ships.
The loop, and why it's the moat
Research → execute → monitor → report. Most tools own one arrow. The value — and the part that's hard to copy — is closing the whole loop in one system.
Here's the version that matters: a strategy gets researched and backtested with walk-forward validation, deployed to a live runtime, run on real capital across multiple exchanges, and monitored with the same attribution and risk rails as everything else on the desk. Not a backtest you trust on faith. Not a black-box signal you can't inspect. A strategy that earned its way from research into live execution and can be watched the whole way.
As AI makes authoring a strategy trivial, the durable value moves off the authoring surface and onto the execution layer — the trusted place where an agent-authored strategy actually runs with real money, safely, with a verifiable live track record. Anyone can get an agent to write a strategy. The hard part is running it. That's the part we built.
What's actually live
We'd rather show you what runs than promise you a roadmap. Today, on the desk:
- A cross-exchange perpetual arbitrage engine runs live on real capital, every day. It watches funding-rate spreads across venues, opens both legs when the net spread clears threshold, manages the position, and closes when the spread compresses. It is the first strategy that proves the runtime works at production scale.
- 28 exchanges connected through one unified surface.
- Non-custodial on funds, always. We never touch your funds. You connect your own exchange keys, scoped to exactly the permissions you grant — and you can run the desk self-hosted, where your keys never leave your machine.
- A research-to-execution pipeline — strategies are backtested and validated before any capital touches them.
We don't publish profit-and-loss numbers, and we don't post trade screenshots. The credibility claim is that the engine exists and runs — not a number we'd ask you to take on faith. The public funding spreads it acts on are verifiable market data; what it earns is ours, and stays private.
What we refuse to be
A thesis is as much what you won't do as what you will.
- We are not a custodian. Non-custodial on funds, always. If that ever changes, it's a different product with a different regulatory posture.
- We are not a data vendor. We integrate best-of-breed data; we don't try to out-build the firms whose whole business is data.
- We are not a “passive income while you sleep” product. The desk is for operators. Risk caps and a kill switch are always on; nothing runs that you didn't author and approve.
- We are not shipping ten things badly. The platform vision is broad; it ships one proven layer at a time. Coherence beats breadth.
Who it's for
Traders who run systematic strategies on their own capital, across more than one exchange. People fluent in funding and basis, past the consumer bot platforms, who want one professional surface instead of a tab stack — and who want the option of an agent operating it alongside them.
If that's you, the list hears first when we open.
Read enough?
Join the waitlistOr keep reading — the engine's running either way. Back to home →