Three ways to deploy
The same environment runs on your servers, in your colocation space, or on dedicated GPU capacity we source and contract for you.
Your servers
We deploy onto GPU servers you already own, in your own facility. Your team keeps physical control. We run the platform and the AI workloads on it.
Your colocation space
We deploy into the space and power you lease from a colocation provider, on hardware you own or on systems we supply.
Dedicated GPU capacity
We source GPU capacity dedicated to you and contract it for you, then deploy and operate the environment on it.
What runs on it
The workloads we deploy and operate inside your environment.
Custom agents on your data
We design and build agents for your own work, then run them in the same private environment.
Multi-agent research
A planner splits each question across specialist agents, and an audit step reviews their work before it reaches you.
Document and regulation extraction
Contracts, filings and rulebooks turned into structured records, each field linked to the passage it came from. See document intelligence.
Internal assistants
Assistants that answer from your documents and act through approved tools, inside your existing access rules.
Grellum, our legal research system, runs this design: a planner, four to six specialist agents and an audit step, with every citation matched to retrieved evidence.
Agentic systemsPlatform and operations
Our own products run on a platform we built and operate this way, with tenant isolation verified by 109 access, admission and network tests.
Kubernetes and GitOps
Every environment is declared in version control and deployed from it, so each change is reviewed and can be rolled back.
Distributed storage
Replicated storage for models, documents and search indexes, with snapshots for recovery.
Network-policy isolation
Each tenant and workload gets only the network paths it needs. All other traffic is denied by default.
109 access, admission and network tests passed for tenant isolation on the Rillor platform, 7 Oct 2026
Monitoring and upgrades
We watch health, capacity and model latency, and apply platform and driver upgrades on a schedule agreed with you.
Security and control
You decide who reaches the environment and what leaves it.
- Data location
- Your data stays in your environment. Models come to the data.
- Isolation
- Each tenant runs in its own isolated space, with its own network rules and credentials.
- Access control
- Role-based access for people, and scoped credentials for agents and tools.
- Audit logs
- Sign-ins, changes and agent tool calls are logged with time, identity and result.
- Terms
- Data handling, retention and access terms are set per engagement.
What the 109 isolation tests check
Access tests confirm that a tenant's credentials reach only its own resources. Admission tests confirm that the platform rejects workloads that break policy. Network tests confirm that traffic between tenants is blocked. All 109 passed on the Rillor platform on 7 Oct 2026.
The GPUs to run it
If you need hardware or capacity, we source it and arrange how you pay for it.
How a deployment runs
One team from the first workload review to daily operation.
- ScopeWe map the workloads, data, users and constraints, and choose a deployment model.
- DesignWe size the hardware, platform, models and access rules, and agree the terms.
- DeployWe build the environment, connect your data, and test models and agents before go-live.
- OperateWe run, monitor and upgrade the environment, with change records you can review.
Notices
AI outputs. AI outputs can be wrong. Important decisions should include human review.