Node.js Integration

AI Guardian works with the openai npm package with zero additional dependencies. Change apiKey and baseURL — everything else is identical.

Installation

npm install openai

Basic Usage

import OpenAI from "openai";

const client = new OpenAI({
  apiKey: "aig_YOUR_API_KEY",
  baseURL: "http://localhost:8000/api/v1/proxy",
});

const response = await client.chat.completions.create({
  model: "gpt-4o",
  messages: [
    { role: "system", content: "You are a helpful assistant." },
    { role: "user", content: "What is the capital of France?" },
  ],
});

console.log(response.choices[0].message.content);

Error Handling

Catch APIError from the openai package and check the error.code field:

import OpenAI, { APIError } from "openai";

const client = new OpenAI({
  apiKey: process.env.AI_GUARDIAN_API_KEY,
  baseURL: process.env.AI_GUARDIAN_BASE_URL,
});

async function safeComplete(userMessage: string): Promise<string> {
  try {
    const response = await client.chat.completions.create({
      model: "gpt-4o",
      messages: [{ role: "user", content: userMessage }],
    });
    return response.choices[0].message.content ?? "";

  } catch (error) {
    if (error instanceof APIError) {
      const body = error.error as Record<string, unknown>;
      const code = (body?.code as string) ?? "unknown";

      if (code === "request_blocked") {
        return `[BLOCKED] Risk score: ${body.risk_score}. Blocked by AI Guardian.`;
      }
      if (code === "queued_for_review") {
        return `[QUEUED] Review ID: ${body.review_item_id}. Pending human review.`;
      }
    }
    throw error;
  }
}

Vercel AI SDK

Using the Vercel AI SDK? Use createOpenAI with a custom baseURL:

import { createOpenAI } from "@ai-sdk/openai";

const guardian = createOpenAI({
  apiKey: process.env.AI_GUARDIAN_API_KEY,
  baseURL: process.env.AI_GUARDIAN_BASE_URL,
});

// Use like any other Vercel AI SDK provider
const model = guardian("gpt-4o");

Environment Variables

Store credentials in .env.local (Next.js) or .env:

# .env.local
AI_GUARDIAN_API_KEY=aig_YOUR_API_KEY
AI_GUARDIAN_BASE_URL=http://localhost:8000/api/v1/proxy