Control Zero is in early beta. We ship daily. Feedback: team@controlzero.ai

SDKs

Add governance in minutes.

Available for Python and Node.js with 11+ framework integrations. Wrap your existing LLM client. Everything else stays the same.

01

Python

Python 3.8+
$ pip install controlzero
Documentation

Quick Start

python
import controlzero
from controlzero.integrations.openai import wrap_openai
import openai
cz = controlzero.init(api_key="cz_live_...")
client = wrap_openai(openai.OpenAI(), cz)
# Every call is governed + audited automatically
response = client.chat.completions.create(
model="gpt-4",
messages=[{"role": "user", "content": prompt}]
)
02

Node.js

Node 18+
$ npm install @controlzero/sdk
Documentation

Quick Start

javascript
import { ControlZero } from '@controlzero/sdk';
import OpenAI from 'openai';
const cz = new ControlZero({ apiKey: 'cz_live_...' });
const client = cz.wrapOpenAI(new OpenAI());
// Every call is governed + audited automatically
const response = await client.chat.completions.create({
model: 'gpt-4',
messages: [{ role: 'user', content: prompt }],
});
03

Go (preview)

Go 1.21+ -- currently in preview
$ go get controlzero.ai/sdk/go
Documentation

Quick Start

go
import controlzero "controlzero.ai/sdk/go"
client, _ := controlzero.NewClient(
controlzero.WithAPIKey("cz_live_..."),
)
defer client.Close()
// Enforce a policy before any action
result, err := client.Enforce("llm:invoke", "model:gpt-4")
if result.Allowed {
// proceed with the call
}

Included in every SDK

All SDKs Include

Policy Enforcement

Local policy evaluation with low latency. Compiled policy engine.

Audit Logging

Every decision recorded with full context. Immutable audit trail.

Real-time Sync

Policy changes propagate to all SDK instances in under 1 second.

Zero Overhead

In-process evaluation. No proxy, no sidecar, no per-call network round trip.

Full documentation at docs.controlzero.ai

Guides, API reference, and examples for every SDK.