agent-callbacks

What problem does it solve?

Agent execution can feel like a black box, making it hard to debug failures, verify tool behavior, and monitor model calls in production.

Core Features & Use Cases

• Model lifecycle callbacks: run logic before and after LLM requests to log inputs, tools available, and response summaries.
• Tool lifecycle callbacks: capture tool start/end events to audit arguments and results.
• Sub-agent delegation hooks: intercept and observe events produced by child agents when using an AgentTool.
• Error handling hooks: intercept model errors to log issues and optionally return a recoverable response.
• Tracing integration (Langfuse): stream events to a tracing backend via callbacks for monitoring and analytics.

Quick Start

Ask: “Enable model and tool callbacks for my orxhestra agent to log requests, tool calls, and errors during streaming runs.”