Frequently asked questions about NEXUS Runtime.

NEXUS Runtime is the reliability control plane for production AI and an AI Runtime Operations Platform. It combines gateway routing, provider model catalogs, AI prompts repo workflows, Reliability Score, root cause analysis, eval gates, policy controls, AI Release Manager releases, AI change management, observability, AI infrastructure management, and deployment handoff.

NEXUS Runtime is the product name for the NEXUS AI Runtime layer. It manages AI runtime traffic, prompts, model routing, observability, releases, governance, and deployment handoff for production AI systems.

The AI prompts repo is the prompt registry where teams version prompts, select real provider-supported models, run tests, enforce eval gates, and promote approved prompt changes across environments.

The AI Release Manager tracks deployment requests, prompt releases, pipeline status, approvals, rollback context, and AI change management evidence before requests are handed to NEXUS AI.

The provider layer includes OpenAI, Anthropic, Google Gemini, AWS Bedrock, Mistral AI, Cohere, Groq, and DeepSeek. Model dropdowns are driven by provider-specific live catalog calls where credentials are configured.

NEXUS Runtime is for engineering, platform, AI operations, security, and product teams that run AI features in production and need routing, governance, release evidence, observability, and reliability controls.

No. NEXUS Runtime manages the production AI runtime layer. NEXUS AI remains the deployment platform with cloud adapters, cloud execution, and cloud credential ownership.

Reliability Score summarizes runtime health signals such as request success, latency, provider errors, failover behavior, policy blocks, and deployment risk into a score teams can use to understand production AI reliability.

Root Cause Analysis reviews runtime signals such as logs, traces, provider errors, policy decisions, prompt changes, model routing, and deployment context to explain why an AI request or workflow failed.

Model lists are provider-specific. NEXUS Runtime fetches supported models from configured provider APIs where possible and limits prompt testing and routing choices to models supported by the selected provider.

No cloud deployment credentials need to live in NEXUS Runtime. For deployments and AI infrastructure management, users add a NEXUS AI access token and NEXUS AI handles the cloud adapters and cloud credential boundary.

A workspace admin adds a NEXUS AI access token created on the NEXUS AI platform. NEXUS Runtime uses that token to send deployment requests to the configured NEXUS AI URL.

AI provider usage is billed by the configured provider account. NEXUS Runtime manages the operational layer for routing, governance, prompt workflows, observability, releases, and reliability controls.

Observability uses runtime request logs, gateway telemetry, traces, provider health, rate limit data, policy decisions, deployment history, prompt versions, agent activity, and Redis or service health where configured.

NEXUS Runtime tracks deployed versions, release context, pipeline state, and rollback candidates so teams can identify the previous approved prompt or deployment state and hand the rollback request to NEXUS AI.

NEXUS Runtime is structured around tenant isolation, scoped API keys, provider credential boundaries, policy controls, audit signals, release history, observability, and NEXUS AI deployment handoff.

The core modules are wired to app data: Prompt Registry, Auth and Rate Limit, Model Router, Observability, and Deployment pages pull real records from backend endpoints.