Building AI applications often feels like assembling a 1,000-piece puzzle where half the pieces are infrastructure boilerplate. You need streaming routes, provider adapters, client instances, and a responsive UI—all before you even write your first prompt.

Today, we're simplifying this with cencori app. This guide walks you through the 60-second path to a running AI application.

Step 1: Initialize your Project

The beauty of create-cencori-app is that you don't need to clone any repositories or manage boilerplate yourself. Just run the following command in your terminal:

Codetext

npx create-cencori-app my-ai-app

Step 2: Choose your Foundation

The CLI will guide you through an interactive setup process. You'll need to make two primary decisions:

Select a Framework:
- Next.js: The recommended choice for full-stack applications. It comes pre-wired with the Vercel AI SDK and App Router.
- TanStack: Choose this for lightweight Vite-based applications using React Query for data fetching.
Include Demo Chat UI?:
- Select Yes to inject a pre-built chat component. This gives you a premium, production-ready interface out of the box.

Step 3: Configure your API Key

Once the project is scaffolded, navigate into the directory:

Codetext

cd my-ai-app

Open the .env.local file (for Next.js) or .env (for TanStack). You'll see a placeholder for your Cencori API key.

Codetext

# Get your key at https://cencori.com/dashboard
CENCORI_API_KEY=csk_your_actual_key_here

[!TIP] If you provided your key during the CLI prompts, this step is already done for you!

Step 4: Launch and Test

Now, start your development server:

Codetext

npm install
npm run dev

Open http://localhost:3000 in your browser. If you opted for the Chat UI, you'll see a minimalist, high-end interface ready for its first message.

Type a question, and watch as Cencori handles the streaming, model routing, and security policies in the background.

Step 5: Advanced Customization

Your project includes a cencori.config.ts file. This is where you can easily swap models or adjust settings without touching your application logic.

To change your default model from GPT-4o to Claude 3.5 Sonnet, simply update the config:

Codetext

// cencori.config.ts
export const cencoriConfig = {
    defaultModel: 'claude-3-5-sonnet',
    // ...
};

Build your AI Future

You now have a production-grade AI infrastructure. From here, you can start building your unique AI experience, leveraging Cencori's unified gateway to access any model with built-in security and failover.

Start building today at cencori.com/docs.