Saltar al contenido

Guardarraíles (Guardrails)

Controles de seguridad alrededor de un sistema de IA que limitan lo que puede decir o hacer: filtran contenido dañino, restringen herramientas y aplican políticas.

Essential for agents that act on real systems: cap loop counts, require approval for irreversible actions, and log every tool call. Managed platforms like Bedrock ship guardrails as a built-in feature.

Recursos relacionados

Amazon Bedrock
A managed AWS service offering foundation models from multiple providers behind one API, with agents, guardrails, and knowledge bases.
LangGraph
A low-level orchestration framework from the LangChain team for building stateful, multi-step agent workflows with fine-grained control.
Dify LLM App Development Platform
An open-source LLM application development platform offering visual orchestration, a RAG engine, and Agent capabilities for building AI apps quickly.