Launching today

AgentDbg

Debug everything your AI Agent does, locally

5 followers

Record agent runs to local JSONL and inspect a timeline of LLM calls, tool calls, state, and errors. Includes loop warnings. No cloud, no accounts, no telemetry. 1. Timeline UI for LLM/tool calls/errors/state 2. Loop warnings for repetitive behavior 3. Redaction on by default + truncation limits 4. Works with any Python; optional LangChain/LangGraph handler 5. Export runs to a single JSON

AgentDbg Reviews

Reviews
Most Informative