What problem does it solve?
Provides secure, centralized access to external LLM providers from within Domino, eliminating the need to distribute provider API keys and giving operators visibility and control over model usage and costs.
Core Features & Use Cases
- Secure LLM Proxy: Routes workspace and job requests to external providers through a Domino-managed gateway.
- Centralized API Key Management: Provider keys are stored in a secure vault with rotation and never exposed to end users.
- Observability & Controls: Token usage metrics, audit logs, rate limits, cost alerts, and per-endpoint access control for compliance and cost management.
- Use Case: Platform teams expose an endpoint (e.g., openai-gpt4) so data scientists can call advanced models from jobs and notebooks without needing provider credentials while admins monitor usage and enforce limits.
Quick Start
Use the Domino AI Gateway endpoint named production-gpt4 to route LLM calls through Domino, enforce access control, and inspect usage metrics for cost and audit purposes.