AgentShield for OpenAI

Track OpenAI API costs per call, monitor token usage, detect anomalies on your GPT-4 and GPT-3.5 agents.

Quick setup

import agentshield
from openai import OpenAI

shield = agentshield.init(api_key="your-key")
client = shield.wrap_openai(OpenAI())

# All calls now tracked automatically
response = client.chat.completions.create(
    model="gpt-4",
    messages=[{"role": "user", "content": prompt}]
)