AI Gateway
Govern every provider request from one enterprise gateway.
NEXUS Runtime manages provider credentials, tenant API keys, rate limits, live provider model catalogs, routing configuration, AI prompts repo testing, and request telemetry for production AI traffic in the AI Runtime Operations Platform.
Supported provider familiesOpenAI, Anthropic, Google Gemini, AWS Bedrock, Mistral AI, Cohere, Groq, and DeepSeek are represented in the provider layer. Model lists are fetched from configured provider APIs when credentials are available.
Gateway capabilities
Tenant API keysIssue scoped keys for gateway access, show prefixes for identification, and enforce key-specific limits.
Auth and rate limitsPull real key, quota, and usage data from the app instead of showing static dashboard values.
Live model catalogFetch supported model IDs from provider APIs so prompt testing and routing use real options.
Model routerUse provider availability, model support, weights, and failover behavior to route requests.
Provider healthSurface circuit state, latency, errors, and last-seen signals from the gateway runtime.
Policy checksApply PII, model restriction, spend limit, and time gate policies before provider execution.