The largest swarms
The highest agent count available, each holding its own exclusive share of the market, so the fleet covers more at once without agents competing.
Being direct about this saves everyone a call.
Not a different product. More of it, wired into yours.
The highest agent count available, each holding its own exclusive share of the market, so the fleet covers more at once without agents competing.
Position ceilings, loss floors and concurrency set to your framework rather than chosen from a dropdown — including the ability to vary guardrails your mandate requires varied.
Scoped API credentials for your own systems: read state, drive controls, pull the full activity record on your own schedule.
Choose which predictive models hold allocation, rather than letting measured performance decide it for you.
Single sign-on through your own identity provider, so joiners and leavers are handled where that already happens.
Not a shared inbox and not a ticket queue. Someone who knows your configuration and answers about it.
Most desks already have risk, reporting and reconciliation they trust. The agent is a source of activity for those systems, not a replacement that asks them to move.
Query state and activity whenever you want it, at whatever frequency your own systems run at.
Every action individually, with result, latency, venue and time — not a daily aggregate you have to take on trust.
Every credential is named and individually revocable, and every use is logged against the key that made it.
And, just as usefully, what we will not.
Non-custodial on every plan, including this one. There is no enterprise arrangement in which we hold your keys, and we will put that in the agreement.
The risk envelope we agree is the envelope the system enforces. If you need that stated as an obligation rather than a setting, we will state it.
Your activity is yours and exports cleanly. We will commit to that in writing, because a retention advantage built on friction is not one worth having.
We will not contract to a return, a hit rate or a Sharpe. Anyone who will is describing something other than autonomous trading, and you should ask them what.
We will not disclose how the agent finds or acts on opportunities, under NDA or otherwise. That is firm, and it applies to every plan.
The risk controls and the custody model are identical on all three. What changes is scale and how far it reaches into your stack.
| Capability | Basic | Pro | Enterprise |
|---|---|---|---|
| Agent swarms | Small | Large | Largest |
| Networks covered | Core | 15+ | 15+ |
| Predictive models | — | Included | Included |
| Choose which models run | — | — | Included |
| Bespoke risk envelope | — | — | Included |
| Programmatic access | — | Included | Included |
| Federated sign-in | — | — | Included |
| Named contact | — | — | Included |
| Non-custodial | Included | Included | Included |
| Full risk controls | Included | Included | Included |
| Immediate stop | Included | Included | Included |
We confirm what you intend to run and whether this fits. Some conversations end here, and that is the purpose of the step.
Limits, roles and stop authority agreed explicitly and written down before anything is configured.
Your envelope is set up with you present, and a wallet you control is connected. Nothing is left on a default to adjust later.
Trading begins inside the agreed envelope, with the full record running from the first fill.
On every plan, without exception.
Set to your framework, not ours.
A person, not a queue.
Exports cleanly, contractually.
“Our requirement was simple and nobody could meet it: cover the venues we cannot staff overnight, inside our risk limits, without us ever handing over the keys.”
We will tell you whether this fits before anyone signs anything, and we will not quote you a return to get the meeting.
Non-custodial on every plan. Trading involves risk of loss.