sentinel-sim 0.2.0

Python SDK for Sentinel SIM — AI agent governance, tool routing, and observability

sentinel-sim

Python SDK for Sentinel SIM — the infrastructure layer for AI agent tool calls.

Route any tool call through Sentinel with full observability, policy enforcement, and budget control.

Install

pip install sentinel-sim

Quick start

from sentinel import Sentinel

client = Sentinel(
    api_key="sk_agent_...",           # from dashboard.sentinel-sim.com
    base_url="https://web-production-90bb3.up.railway.app",
)

# Chat via OpenAI
result = client.openai_chat([
    {"role": "user", "content": "Hello!"}
])

# Deploy to Vercel
result = client.vercel_deploy("my-app", files=[
    {"file": "index.html", "data": "<h1>Hello</h1>"}
])

# Create a GitHub repo
result = client.github_create_repo("my-repo", private=True)

# Deploy a Railway service
result = client.railway_create_project("my-project")

Async support

from sentinel import AsyncSentinel
import asyncio

async def main():
    async with AsyncSentinel(api_key="sk_agent_...") as client:
        result = await client.openai_chat([{"role": "user", "content": "Hi"}])

asyncio.run(main())

Supported providers

Provider	Tools
OpenAI	chat, embedding, image_generation, transcription
Anthropic	chat, vision, tool_use, streaming
GitHub	create_repo, push_file, create_issue, create_pr
Vercel	deploy, manage_project, manage_domain
Railway	create_project, create_service, deploy_service, upsert_variables
Stripe	create_payment_intent, create_customer, create_refund
Replicate	image_generation, model inference
Resend	send_email

Links

• Dashboard
• API Docs