04
AI governance & operations
Every model call passes through one gate: routing, budgets, quotas and audit.
The LLM Gateway is the single entry point for text generation and embeddings. Features cannot connect to model providers on their own; every call is authorized, routed, budgeted and audited.
- 04.1
Provider registry
Admins register approved model deployments; unregistered endpoints cannot be used.
- 04.2
Routing
Requests are routed to the right deployment by task and policy.
- 04.3
Budgets & quotas
Per-department budgets and quotas, handled by rule when exceeded.
- 04.4
Policy & audit
Personal-data and cloud-use policies are enforced at the gate, and every call is logged.
One path, fully traceable
- Meetings
- Knowledge
- Assistant
LLM Gateway
- Auth
- Routing
- Budget / quota
- Policy
- Audit
Model deployments you approve
- LLM Gateway
- Authorization, routing, budgets and audit in one place.
- Department quotas
- Allocate usage by department.
- Prompt management
- Central prompts and templates, with prompt-injection detection.
- Audit logs & API keys
- Traceable operations; API keys issued and revoked centrally.
- Support center & notifications
- Built-in tickets and notification preferences.
Which model deployments are available is decided by your environment and policy. This site does not name specific providers.