Skip to content
Cadence AI Orchestration Platform

Feature guides

In-depth guides for every major Cadence capability — concepts, workflows, code, and troubleshooting.

Intended audience: Stakeholders, Business analysts, Solution architects, Developers, Testers

Learning outcomes by role

Stakeholders

  • Navigate feature areas to prioritize roadmap and risk conversations.

Business analysts

  • Map product epics to the feature guides listed in this section.

Solution architects

  • Choose deep dives for integration design (tenancy, auth, orchestration, plugins).

Developers

  • Jump to API- and code-oriented pages for each subsystem.

Testers

  • Plan coverage by feature domain using the sidebar structure.

In-depth guides for every major Cadence capability. Each page leads with who it is for, a visual or diagram where it helps, and technical detail with repository paths—plus limitations and related pages at the end.

If you are new to the platform, start with What is Cadence? and How the platform works, then use the sections below.

Security and tenancy

Section titled “Security and tenancy”
Multi-tenancy Organization isolation, X-ORG-ID context, and request lifecycle.
Role-based access control cadence:* permission strings, built-in roles, and API key scopes.
JWT sessions Redis-backed jti sessions, revocation, and API key contrast.

LLM and orchestration

Section titled “LLM and orchestration”
LLM configuration BYOK provider setup, encrypted API key storage, and framework compatibility.
Orchestration backends LangGraph and OpenAI Agents SDK — framework capabilities, modes, and AI App creation.
Orchestration modes Supervisor vs Grounded, resource_id, and mode_config.
AI Apps Create, load, and manage AI Apps per org.
AI Agent system Upload-attach-load lifecycle for ZIP plugin packages (AI agents).
AI Agent SDK BasePlugin, BaseAgent, UvTool, and the full SDK reference for plugin authors.
AI Agents catalog System and org plugin catalog, upload validation, and settings schema.
Hot-reload and AI App pool Pool tiers, async load/unload, and 202 semantics.

Chat and observability

Section titled “Chat and observability”
Chat and engine Chat completion headers, AI App vs central point routing, and engine prompts.
Real-time streaming SSE event model, header contract, and client parsing.
Central Points Stable aliases for AI Apps with private/public visibility.
Observability Structured logs, health endpoints, and OpenTelemetry.