# Candu Actions

## Docs

- [Actions](https://agents.candu.ai/actions/actions.md): Define what your AI agent can do. Each action maps to one real API endpoint.
- [How Candu Actions works](https://agents.candu.ai/actions/architecture.md): The three-layer architecture and how your app, Candu, and your API fit together.
- [Create a segment](https://agents.candu.ai/actions/create-segment.md): A complete walkthrough from user input to confirmed creation and validation. 6 API calls, 2 wizard steps, 1,033ms total.
- [Discovering actions](https://agents.candu.ai/actions/discovery.md): Three ways to populate your action registry from your existing product.
- [Execution & Logs](https://agents.candu.ai/actions/execution.md): Every action is logged. When something fails, Candu Actions tells you exactly which layer broke and proposes a fix.
- [Getting Started](https://agents.candu.ai/actions/getting-started.md): One snippet. No backend. First action running in under 10 minutes.
- [Turn user intent into real product actions](https://agents.candu.ai/actions/overview.md): Candu Actions doesn't generate answers. It executes them. Give your AI a structured, safe, auditable way to call your API.
- [Policies](https://agents.candu.ai/actions/policies.md): Control how actions execute, who can trigger them, and when confirmation is required.
- [Response Mask](https://agents.candu.ai/actions/response-mask.md): Control which fields from a tool's HTTP response are forwarded to the AI model in subsequent turns.
- [Wizards](https://agents.candu.ai/actions/wizards.md): Structured, step-by-step flows for guided product workflows.

## OpenAPI Specs

- [openapi](https://agents.candu.ai/api-reference/openapi.json)