Developer Platform

Runtime governance for autonomous AI agents. Intercept every tool call. Enforce policies. Detect threats. Ship safely.

Govern your first tool call

Install an SDK or use cURL directly. Every tool call evaluation returns a decision (allow/review/block), a risk score, and a full explanation.

from airg import GovernorClient

client = GovernorClient(
    base_url="https://api.airg.nov-tia.com",
    api_key="airg_your_key_here"
)

result = client.evaluate(
    tool="shell",
    args={"command": "rm -rf /"},
    context={"agent_id": "my-agent"}
)

print(result.decision)   # "block"
print(result.risk_score) # 95

Everything you need

Official client libraries