# Extra

## Docs

- [HTTP API](https://docs.extra-ai.co/docs/api.md): REST endpoints for the engine and the conversation layer.
- [Architecture](https://docs.extra-ai.co/docs/architecture.md): How Extra works — from YAML to a running response.
- [CLI Reference](https://docs.extra-ai.co/docs/cli.md): All agentctl commands and their flags.
- [Execution Limits](https://docs.extra-ai.co/docs/execution-limits.md): Built-in guardrails that cap tool calls and iteration loops per run.
- [Introduction](https://docs.extra-ai.co/docs/introduction.md): Extra — an agent engine for hierarchical multi-agent systems.
- [MCP & Tools](https://docs.extra-ai.co/docs/mcp-and-tools.md): Connect remote MCP servers and local Python tools to your agents.
- [Observability](https://docs.extra-ai.co/docs/observability.md): Structured logs by default, and Langfuse tracing with two environment variables.
- [Plugins](https://docs.extra-ai.co/docs/plugins.md): Resolver plugins, the access plugin, and how to generate and implement them.
- [Prompt Rendering](https://docs.extra-ai.co/docs/prompt-rendering.md): How prompt files work, how variables get filled, and when rendering happens.
- [Quickstart](https://docs.extra-ai.co/docs/quickstart.md): Get Extra running in minutes — engine only, or with conversation history and an embeddable widget.
- [Roadmap](https://docs.extra-ai.co/docs/roadmap.md): What's implemented, what's coming next.
- [Runtime Hooks](https://docs.extra-ai.co/docs/runtime-hooks.md): Trusted Python code the engine runs automatically at fixed lifecycle points — for auth, validation, policy, and audit.
- [Full Example — Zero to Hero](https://docs.extra-ai.co/docs/tutorial.md): Build a real support system from an empty file, one YAML section at a time.
- [Embeddable Widget](https://docs.extra-ai.co/docs/widget.md): Add a chat widget to any web app with a single script tag.
- [YAML Reference](https://docs.extra-ai.co/docs/yaml-spec.md): Every field you can declare in agents.yml, with examples across different domains.