01
Model routing
Direct requests by policy, workload, availability, cost, or model capability while keeping applications independent from any single provider.
LabOne37 / Product 01 / In development
Odin.Control every model.
An enterprise LLM gateway in development to route models, enforce security, govern access, and turn every token into an observable operational signal.
Route / Models intelligentlyProtect / Every requestObserve / Every token
01 / Gateway architecture
Every query monitored before it reaches a model.
Watch Odin receive traffic from different applications, authenticate the caller, inspect security and PII policy, monitor and optimize token use, then route the approved query to an appropriate LLM endpoint.
Live request trace / illustrative workflowOdin monitoring
Query sources
Internal copilot
HR workflow
Agent workflow
Odin control plane
OdinLLM Gateway / Policy Engine
REQ 1042Monitored query
Waiting for the next application query...
01 / Authentication + SSOAwaiting identity
02 / Security guardrailsAwaiting inspection
03 / Token intelligenceAwaiting context
04 / Model routingAwaiting policy
Token monitor / optimized contextWaiting
LLM model endpoints
Frontier model
Open model
Private model
Odin is ready to monitor the next query.
02 / Capabilities
Governance without slowing down delivery.
Odin centralizes the controls teams otherwise rebuild across applications, providers, and environments.
02
Security guardrails
Inspect prompts and responses, enforce content and data-handling policies, and apply predefined security templates consistently across AI workloads.
03
SSO and access control
Integrate enterprise identity so access to models, policies, environments, and administrative functions follows organizational controls.
04
Token intelligence
Measure token utilization across teams, applications, providers, and models, then optimize context and routing to reduce unnecessary consumption.
05
Logs and dashboards
Trace gateway activity with detailed operational logs and dashboards built for usage analysis, troubleshooting, governance, and review.
06
Alerts
Generate actionable alerts for policy events, anomalous usage, availability changes, budget thresholds, and operational conditions.
03 / Operating model
Security, efficiency, and evidence in the request path.
Identity and policy
Authenticate users and workloads, apply security templates, validate requests, and select eligible model routes.
Routing and optimization
Choose the appropriate endpoint, manage token use, and preserve consistent controls across providers.
Observability and response
Capture detailed events, update dashboards, surface anomalies, and generate alerts for the teams responsible.
04 / Start with Odin
Put one governed gateway in front of your AI estate.
Odin is LabOne37's first product. Contact the team to discuss deployment, integration requirements, and enterprise evaluation.
odin@labone37.com