Every Agent Call.
Logged. SOC 2 Ready.

A lightweight Python middleware that wraps your agent framework and logs every tool call — structured, timestamped, and ready for compliance reviews.

Start Your Free Trial →
⚡ 2 lines of code 🔒 SOC 2 compliant 📄 PDF audit exports 📊 Google Sheets sync

Audit trails your investors will ask for

When you deploy AI agents in production, every tool call is a potential liability. Regulators, compliance officers, and investors want to see exactly what your agents did, when, and with what data.

  • Automatic capture of every tool input, output, and error
  • Structured JSON logs on disk, ready for ingestion
  • One-click PDF audit reports for compliance reviews
  • Google Sheets export for your compliance team
  • Zero impact on agent performance — runs as middleware
{
"id": "a1b2c3d4",
"timestamp": "2026-05-26T14:32:10",
"tool_name": "web_search",
"inputs": { "query": "..." },
"outputs": "...",
"duration_ms": 142.3,
"error": null
}
"error": "RateLimitError: API quota exceeded"
// Every success and every failure. Captured.

Everything you need for agent compliance

From middleware to audit report — in minutes, not weeks.

🔍

Tool Call Logging

Every tool invocation is automatically captured with inputs, outputs, duration, and error state. Decorator-based — wraps any function.

🔄

Data Flow Tracking

Trace data through your agent pipeline. See exactly what was sent to which tool, and what came back. Session-level grouping.

📄

PDF Audit Exports

Generate beautiful PDF reports with reportlab. Session summaries, error breakdowns, and full call logs — ready for compliance reviews.

🔒

Compliance Ready

Structured logging that maps to SOC 2, ISO 27001, and GDPR audit requirements. Every record has a unique ID and immutable timestamp.

📊

Google Sheets Sync

Push logs directly to Google Sheets for your compliance team. No database setup, no infrastructure — just a service account key.

CLI & Export

Built-in CLI for viewing, filtering, and exporting logs to JSON or CSV. Pipe into your existing analytics pipeline in seconds.

Two lines. That's it.

# pip install agent-audit-logger
 
from agent_audit_logger import AuditMiddleware
 
audit = AuditMiddleware("prod-deploy-1")
 
# Your tool, now audited
@audit.wrap("web_search")
def search(query: str):
return tavily.search(query)
 
# Generate a compliance PDF
PDFReport().add_session(audit.to_dict()).save("audit.pdf")

Simple pricing. No surprises.

Start free, scale when you need more. Cancel anytime.

Starter

£49/mo
billed monthly
  • Up to 3 agents
  • 10,000 log records / month
  • JSON & CSV export
  • CLI tools
  • Email support
Get Started

Start logging your agents today

Get early access. No credit card required. We'll send you the package, docs, and a sandbox to try it out.

✉️ Or join our mailing list — no spam, just product updates.