With your permission, we use Google Analytics cookies to count visits and see how people find this site and which pages they read. They are not used for advertising. You can change your choice at any time from Cookie settings at the bottom of every page. Privacy notice
It is software. Its execution engine connects to your exchange through an API, and it runs on whatever hardware you point it at — a Mac mini in your office, a server in a colocation facility, or a virtual private server in the cloud. Or we run it for you.
01
Three ways to host it yourself.
It is lightweight. The research and risk run centrally; your AI receives the output and executes. It does not need a powerful machine. It needs to be on and connected.
Mac mini
Run it in your office.
The simplest path. Plug in a Mac mini, paste in your exchange API keys, and your AI runs continuously in the background. No IT team required.
Any current M-series Mac mini is sufficient
macOS or Linux; we set it up with you during onboarding
Runs on a standard office connection
Remote access via SSH or screen sharing if you need to check it
Dedicated server
Your data room or colocation.
For institutional desks. Your AI runs on your own server hardware in your data room or a colocation facility. Full control over the hardware, network, and security perimeter.
Any Linux server — single blade, colocation rack, or on-premise hardware
Runs alongside your existing trading infrastructure
Connect to your existing monitoring and alerting stack
Suits larger books and desks with an existing colocation arrangement
Virtual private server
Cloud infrastructure.
Run your AI on a VPS at AWS, Hetzner, DigitalOcean, or your existing cloud account. No hardware to manage. Spin up in minutes. Shut it down if you stop.
An entry-level private server is enough
Ubuntu LTS recommended; we provide the deployment steps
Cloud-native monitoring and uptime guarantees from your chosen provider
Easy to migrate between providers if you want to change your setup
02
Or we run it for you.
If you would rather not manage hosting, we run your AI on our infrastructure, dedicated to you. Nothing to install or maintain.
Managed infrastructure
Runs on our infrastructure, dedicated to you and shared with no other client
You provide the exchange API key once. Your AI handles everything from there.
Daily CIO and trader reports delivered to your inbox
Monitoring, uptime, and failover managed by us
Same API key restrictions apply: read and trade only, no withdrawal access
Managed hosting is included with the annual subscription.
03
Technical requirements.
Your AI is intentionally lightweight. The research and risk computation runs centrally. Its job is to receive that output and trade your exchange accounts through its execution engine.
01
Operating system.
macOS, Linux, or Windows Server. On a server or cloud machine we recommend Linux (Ubuntu LTS); on a Mac mini, macOS.
02
Network.
A stable internet connection is all that is required. The execution engine connects outbound to exchange APIs. No inbound ports are needed. Latency to the exchange matters less than uptime — the strategies are not latency-sensitive at the millisecond level.
03
API keys.
You create a read and trade API key on your exchange account and add it to your AI during setup. No withdrawal permission is required or requested. You can revoke the key at any time from your exchange dashboard.
04
Uptime.
Your AI receives risk updates and news in real time, so it should stay on and connected. If a self-hosted AI goes offline, it places no new orders, and we alert you. For larger books, a colocated or cloud server is the more resilient choice.
05
Resources.
Your AI is not computationally intensive. The heavy research runs centrally; yours only receives the output. A Mac mini, an entry-level private server, or a single blade in a shared rack is sufficient.