One decorator. That's the whole integration.

$ pip install foxy-audit

from foxy_audit import audit

@audit(policy="soc2")
def ask_model(prompt):
    return llm.generate(prompt)

The decorator wraps your function transparently. In the default best-effort mode, your code returns immediately while Foxy records what happened in the background. Critical services can enable audit_required=True to fail closed when evidence cannot be delivered.

What actually leaves your machine

• A customer-keyed HMAC commitment of the prompt and response — never the raw text.
• The token count, model name, timestamp, and the policy tag you set.
• The policy verdict from the configured judge — clean, flagged, or unknown.

Built to stay out of your way

• Two delivery modes: best-effort background capture or audit_required fail-closed capture.
• Default mode is best-effort: telemetry errors stay out of your app; audit_required intentionally fails closed.
• Survives outages: records queue locally and sync when the backend is reachable again.