Agents

The following section shows you how to build agents with the AI SDK - systems where large language models (LLMs) use tools in a loop to accomplish tasks.

• Overview: Learn what agents are, how they manage state, and why to use the ToolLoopAgent.
• Building Agents: Create ToolLoopAgent instances with tools, runtime context, and loop control.
• Workflow Patterns: Structured patterns using core functions for complex workflows.
• Loop Control: Advanced execution control with stopWhen and prepareStep.
• Configuring Call Options: Pass type-safe runtime inputs to dynamically configure agent behavior.
• Tool Approvals: Review, approve, or deny sensitive tool calls before they execute.
• Policy-Based Tool Approvals: Author tool authorization rules as code with @ai-sdk/policy-opa and Open Policy Agent.
• Subagents: Delegate context-heavy tasks to specialized subagents while keeping the main agent focused.
• WorkflowAgent: Build durable, resumable agents with @ai-sdk/workflow for the Workflow SDK.
• Terminal UI: Run ToolLoopAgent instances in an interactive terminal interface.
• Chat Platforms: Run agents in Slack, Microsoft Teams, Discord, and more with Chat SDK.