Copy page

Getting Started

Integrations and SDK

Connect Ice Phi to your AI pipelines using our official Python PIP package.

Python PIP Package

Target Runtimes: Python 3.8+ | FastAPI | LangChain | LlamaIndex

Status: Published on PyPI

https://pypi.org/project/icephi/

Installation

Code
pip install icephi

Quickstart Example

Code
from icephi import IcePhi

client = IcePhi(api_key="your_api_key")

response = client.scan(prompt="User input string here")

if response.is_flagged:
    raise ValueError("Prompt injection threat detected.")

TypeScript SDK (@ice_phi/icephi-ts)

Target Runtimes: Node.js 18+ | Express | Fastify | Custom HTTP Loops

Installation

Code
npm install @ice_phi/icephi-ts

Quickstart (Express Middleware Example)

Code
import express from 'express';
import { IcePhiShield } from '@ice_phi/icephi-ts';

const app = express();
app.use(express.json());

const shield = new IcePhiShield({ apiKey: process.env.ICEPHI_API_KEY });

// Option 1: Mount it explicitly to a /shield prefix path
app.use('/shield', shield.middleware());

app.listen(3000);

Zero-Data-Retention Policy

Payload content sent to Ice Phi endpoints is processed strictly in-memory during real-time threat inspection. No prompt inputs or model outputs are written to disk, logged, or cached.

For enterprise compliance questions or support, contact [email protected].