I built agent-inspect to debug TypeScript AI agent trajectories
Your AI agent failed. Again. The final answer is wrong, but the logs look fine: tool call started model call started tool call completed model call completed fallback used error: timeout Which tool caused the timeout? Did the model answer before retrieval finished? Was the fallback expected? Did the agent call the same tool twice? This is where console.log stops feeling like debugging and starts…
I built agent-inspect to debug TypeScript AI agent trajectories. When an AI agent fails, the logs often appear fine, but the tool calls, model calls, fallback usage, and timeouts can be difficult to pinpoint. AgentInspect addresses this issue by transforming a local trace into three valuable outputs: a readable execution tree, a deterministic regression gate, and an Evidence v2 bundle for review.
The tool provides a readable execution tree, a deterministic regression gate for testing, and an Evidence v2 bundle for sharing. It focuses on metadata by default, with no accounts, collectors, or default uploads. The debugging process involves generating a synthetic demo, initializing the agent-inspect tool, and using it to view, report, explain, and analyze the run.
This approach restores the structure lost in flat logs, showing nested steps, tool and model calls, durations, metadata, errors, and observed outcomes.
Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
