SaolixGroup
Saolix SaoHub. Governed AI agents at enterprise scale.
The Saolix Enterprise Agentic AI Platform. SaoHub is being built to host, run and govern AI agents across your organisation: a marketplace of agents built for specific jobs, a manager to run them or build your own, and a governance layer over every agent.
Pre-release SaoHub is being built. The simulator and diagrams show the planned design, with fictional data; Raia, the first agent, already works inside Saolix Eagle.
Designed so that every agent is registered, every action is governed, and every decision is owned by a rule or a person.
AI agents can gather information, use tools and propose the next step on their own. As organisations add more of them, the hard question stops being whether an agent can act and becomes who authorised it. SaoHub is being built to answer that in one place: agents are registered and permissioned, their actions pass through approved rules and people, and every step is recorded so it can be audited later.
Agents advise. Rules and people decide.
SaoHub is designed on the same principle as every Saolix product: an agent can propose and explain, but only approved rules and people authorise a consequential action.
What SaoHub is being built to do.
Agent Marketplace
Will offer governed agents built for specific jobs in regulated work, such as screening, identity, fraud and security, each with its guardrails and audit built in.
-aia agents · Raia firstAgent Manager
Will let you run the marketplace agents or build your own, and register, permission, monitor and switch off every agent in one place.
run · build · switch offGovernance layer
Will sit over every agent: policies set what an agent may do, approvals are required where your rules say so, and every action is recorded for audit.
policies · approvals · auditPrivate by design
Designed to run inside your environment on locally hosted models, including Sōli models when they are released, with no data sent to an external AI service.
on-premise · local modelsThe agents, and where they stand.
SaoHub agents carry names in the -aia family. Raia is the first, and it already works inside Saolix Eagle.
Raia
Through Eagle's triage API, suggests a label for a screening alert, likely noise or needing review, with a confidence and a reason. It runs on a locally hosted open model. A guard in Eagle overrides any "noise" label on a real hit, and Raia can never close an alert.
Fraud investigation agent
Designed to link related activity across customers, accounts and devices and explain why a pattern looks like fraud. Governed rules decide whether to hold, block or allow.
Threat-hunting agent
Designed to correlate security events into a likely attack path, propose hunts and suggest a response. The response runs only under approved policy, with approvers where the policy requires them.
What no SaoHub agent can do.
Decide
Verdicts, holds and responses come from versioned rules or approved policy, and from people. An agent's output is advisory.
Override a real hit
Where an agent's suggestion conflicts with the rules, the rules win, and the case goes to a person.
Become essential
Switching an agent off removes the help, not the correctness. Every decision still comes out the same.
Designed for your infrastructure.
SaoHub is designed to follow the same deployment model as Eagle: inside your environment, under your control.
Target design of a pre-release product. It shows how SaoHub is designed to work, not a released product. Dashed boxes are planned or outside services.
Explore the design. Inspect it.
SaoHub is in pre-release. The simulator shows how it is designed to work, and its governance pattern already runs in Saolix Eagle.
SaoHub Simulator
Add an agent from the marketplace, see what it may never do, let it propose, decide as two different people, switch it off and read the record.
Open the simulator ↗ Working todayRaia in the Eagle Simulator
See an AI label an alert and the rules decide, in Eagle's interactive simulator.
Open the simulator ↗ BriefingDesign walkthrough
How SaoHub would govern the agents you plan to use, and where the determinism boundary sits.
Request a briefing →Further reading.
Your next chapter.
Let’s build it together.
Explore a product, discuss a project, or find the right expertise. Tell us where you want to go. We’ll help you take the next step.