Deployment model
Cloud (SaaS)
Use the managed lane when rapid rollout, centralized updates, and shared cloud operations matter more than owning the full infrastructure stack.
Deployment
Deployment Guide
Evaluate where the platform runs, how the data plane is bounded, and which operating controls stay inside your environment.
Deployment lanes
Keep the operator surface consistent while you decide where compute, storage, and identity boundaries belong.
Supported deployment lanes
4 models
Core data plane
Postgres + Neo4j
Object storage compatibility
R2 / S3
Identity integration posture
SSO-ready
Architecture brief
Choose the boundary that matches your residency, accreditation, and operating model
Deployment planning should clarify where the data sits, who operates the environment, and how observability, storage, and identity controls are partitioned before implementation starts.
Deployment model
On-Premise
Keep the full stack inside your controlled network when data residency, internal accreditation, or disconnected operations require it.
Deployment model
Hybrid
Split storage, identity, or AI processing across environments when the operational boundary is not the same as the compute boundary.
Deployment model
Government Cloud
Use a government-approved cloud lane when procurement or sovereign hosting requirements call for a controlled public-cloud footprint.
Operator surface
The analyst-facing workflow stays consistent while the underlying residency and hosting model changes.
System Architecture
Core stack
The platform keeps the operator workflow stable while the hosting, storage, and identity controls are mapped to the target environment.
Web Frontend
Next.js / Cloudflare Workers
API Gateway
REST, GraphQL, and event services
AI Services
Model routing and governed processing
Object Storage
Object storage with evidence retention controls
Capacity planning
System Requirements
Infrastructure sizing should reflect operator concurrency, ingestion load, evidence retention, and the models you expect to run in the environment.
Baseline
Minimum (Small Deployments)
Use this lane for pilots, low-volume teams, or controlled proof-of-value environments.
- 8 CPU cores
- 32 GB RAM
- 500 GB SSD storage
- 1 Gbps network
Operational target
Recommended (Enterprise)
Plan for this lane when multiple teams, heavier ingestion, or sustained analytics workloads are expected.
- 32 CPU cores
- 128 GB RAM
- 2 TB NVMe storage
- 10 Gbps network
Planning notes
What usually changes the size
The main drivers are document volume, evidence retention, graph depth, alerting frequency, and how much AI processing stays inside the boundary.
Telemetry
Deployment planning should include the telemetry needed to watch usage, queue depth, and operational load after go-live.
Need an architecture review?
Walk through residency, storage, identity, and observability assumptions with the team before deployment design is locked.