AuditTrail

Don't just trace what happened.
Understand why.

Open-source explainable observability for AI agent systems. Causal attribution + constitutional governance + visual debugging.

AuditTrail DAG Visualization
57%
of orgs have AI agents
45 > 5 min
debugging time reduction
<5ms
middleware overhead
Aug 2026
EU AI Act enforcement

Existing tools show WHAT happened.
AuditTrail shows WHY.

ToolCausalGovernanceSankeyOpen Source
LangSmithNoNoNoNo
LangfuseNoNoNoYes
Fiddler AINoYesNoNo
BraintrustNoNoNoNo
HeliconeNoNoNoYes
AgentOpsNoNoNoNo
AuditTrailYesYesYesYes
Explore the Platform

Compliance Dashboard

Constitutional governance with real-time amber/red boundary detection. Define safety rules in YAML. Monitor compliance across all agents.

Compliance DashboardCompliance Dashboard
Sankey AttributionSankey Attribution
Trace Detail DAGTrace Detail DAG
Analytics DashboardAnalytics Dashboard
Timeline WaterfallTimeline Waterfall

Two Lenses. One Platform.

Same trace data, two powerful perspectives for understanding agent behavior.

Sankey View

Sankey Causal Attribution Diagram

Causal attribution from prompt phrases through reasoning steps to tool calls.

DAG View

DAG Decision Chain Visualization

Interactive decision-chain visualization with constitutional compliance overlay.

Get Started in Seconds

Three lines of code. Full explainability. No configuration needed.

agent.py
1 import audittrail
2
3 # Initialize -- one line, zero config
4 audittrail.init(frameworks=["langgraph"])
5
6 @audittrail.traceable
7 async def run_agent(prompt: str):
8 result = await graph.ainvoke({"input": prompt})
9 return result
10 # Full traces, DAG, Sankey -- automatic
Terminal
$ docker-compose up -d
$ open http://localhost:3000
# Dashboard ready. Start tracing.

1. Install the SDK

pip install audittrail -- works with LangGraph, CrewAI, and raw OpenAI SDK.

2. Launch the Dashboard

One command. Full observability UI at localhost:3000. No cloud. No config.

Ready to understand your agents?

Everything you need to understand your agents

Six core capabilities that transform how you debug, monitor, and govern AI agent systems.

Interactive DAG

Decision chain visualization with collapse, zoom, and minimap

Sankey Attribution

Prompt-to-tool causal mapping with attribution scores

Constitutional Governor

YAML safety rules with amber boundary detection

Real-time Streaming

Watch agent execution unfold live via WebSocket

SHAP Explanations

Feature importance for tool selection decisions

PDF Audit Reports

One-click compliance reports for external auditors