Skip to content

AgentDbg

AgentDbg is a local-first debugger for AI agents. It captures structured traces (LLM calls, tool calls, state, errors) and gives you a timeline UI to inspect what happened-inputs, outputs, latency, and loop warnings.

What it is: A developer tool to instrument your agent, run it, and see a full event timeline locally. No cloud, no accounts.

What it is not: It is not observability or production monitoring. It does not do deterministic replay (planned for a later version), and it does not lock you into any framework.


In 60 seconds

1. Install:

pip install agentdbg

Or from source:

git clone https://github.com/AgentDbg/AgentDbg.git
cd AgentDbg/agentdbg
uv sync

2. Run the example agent:

python examples/minimal/simple_agent.py

3. Open the timeline viewer:

agentdbg view

A browser tab opens showing every event in the run - tool calls, LLM calls, timing. Data is stored locally under ~/.agentdbg/runs/<run_id>/.


Demos and examples

Example Path How to run
Minimal agent (pure Python) examples/minimal/ python examples/minimal/simple_agent.py
LangChain minimal examples/langchain/minimal.py uv run --extra langchain python examples/langchain/minimal.py
OpenAI Agents minimal examples/openai_agents/minimal.py uv run --extra openai python examples/openai_agents/minimal.py
LangChain customer support (advanced) examples/langchain/ Set API keys, then follow _customer_support/README.md
Demos (short scripts) examples/demo/ python examples/demo/pure_python.py or python examples/demo/langchain.py

After any run, open the timeline with agentdbg view.


Documentation

Page Description
Getting started Installation (uv/pip), quickstart, data dir, redaction
Guardrails Stop runaway runs with loop, count, and duration limits
Regression testing Baseline, assert, and diff workflow for catching agent regressions
CLI list, view, export, baseline, assert, diff with options and exit codes
Viewer Timeline UI usage, URL params, live refresh, and development
SDK @trace, traced_run, has_active_run, record_llm_call, record_tool_call, record_state
Integrations LangChain handler, OpenAI Agents adapter, and planned adapters
Architecture Event schema, storage layout, viewer API, loop detection
Reference
Trace format Event envelope, event types, payload schemas, run.json (public contract)
Configuration Env vars, YAML precedence, redaction, truncation, loop detection, guardrails
Policy YAML Assertion policy file format, fields, threshold semantics, CLI mapping