AlixWang

Posted on Nov 6

Building an AI Cron Builder with Cloudflare Pages and Next.js

#ai #cloudflare #pages #nextjs

This technical document outlines how to build an AI-powered cron expression builder using Cloudflare Pages, Next.js, and Cloudflare's AI capabilities. The result site is cronbuilder.online

Project Overview
Technical Stack
Setup and Configuration
Key Components
AI Integration
Deployment

1. Project Overview

The AI Cron Builder is a web application that helps users create cron expressions through:

Natural language input processing using AI
Visual cron expression builder
Real-time validation

2. Technical Stack

Frontend Framework: Next.js 14
Hosting: Cloudflare Pages
AI Model: Cloudflare Workers AI (Llama 3)
UI Components: Ant Design
Styling: Tailwind CSS

3. Setup and Configuration

3.1 Project Structure

app/
├── api/
│   └── cron/         # AI endpoint
├── components/       # React components
├── layout.tsx       # Root layout
└── page.tsx         # Main page

3.2 Edge Runtime Configuration

Enable edge runtime in your API routes by adding:

export const runtime = 'edge'

3.3 Cloudflare Configuration

Create a Cloudflare Pages project
Enable Cloudflare AI in your project settings
Configure environment variables for AI access

4. Key Components

4.1 Main Page Layout

The main page layout (referenced from app/page.tsx) includes:

export default function Home() {
  return (
    <main className="flex min-h-screen flex-col items-center px-12 bg-[#FAFDFF]">
      <header className="w-full h-16">
        <nav className="w-full h-full">
          <ul className="flex list-none h-full items-center justify-between">
            <li className="flex self-center items-center">
              <Image alt="logo" className="w-10 h-10 mr-2" width={100} height={100} src={"/logo.png"} />
              <h1 className="text-lg font-bold text-[#44779C]">Cron Builder</h1>
            </li>
            <li>
              <a role="button" tabIndex={0} href="mailto:service@cronbuilder.online" className="px-2.5 py-1 bg-blue-400 font-bold text-base text-white rounded-md">
                Contact
              </a>
            </li>
          </ul>
        </nav>
      </header>
      <h1 className="text-clip bg-clip-text text-transparent bg-[length:200%_100%] bg-ai animate-ai-animate-btn font-bold text-5xl text-center mt-10 ">
          Build Your Perfect
          <br/>
          Cron Expression Effortlessly
      </h1>
      <section className="w-full mt-8 flex justify-center">
        {/* <h2>Generate your cron expression</h2> */}
        <CronComp />
      </section>
      <section>

      </section>
    </main>
  );
}

4.2 Cron Component

The core cron builder component (referenced from app/components/cron/index.tsx) provides:

Visual cron expression builder
Expression validation
AI integration

const CronComp = () => {
    const [value, setValue] = useState('* * * * *');
    const [error, setError] = useState<boolean>(false);
    const handleCatchError = (error: CronError) => {
        if(error) {
            setError(true)
        } else {
            setError(false)
        }
    }

    return <ConfigProvider
        theme={
            {
                token: {
                    borderRadius: 0
                }
            }
        }
    >
        <div className='flex flex-col min-w-full'>
            <div>
                <AiGen onGenerate={(string) => {
                    if(string) {
                        setValue(string)
                    }
                }} />
            </div>
            <div>
                <Cron onError={handleCatchError} locale={NO_PREFIX_SUFFIX_LOCALE} value={value} setValue={setValue} />
            </div>
            <div>
                <Input status={error ? 'error' : ''} onChange={(e) => setValue(e.target.value)} value={value} />
            </div>
        </div>
    </ConfigProvider>
}

4.3 AI Generation Component

The AI generation component (referenced from app/components/AiGen/index.tsx) handles:

User prompt input
AI request handling
Response processing

const AiGen:FC<AiGenProps> = ({onGenerate}) => {

    const [prompt, setPrompt] = useState<string>();
    const [loading, setLoading] = useState<boolean>(false);

    const btnClick = () => {
        if(loading) {
            return
        }
        setLoading(true);
        fetch('/api/cron', {
            method: 'POST',
            headers: {
          'Content-Type': 'application/json',
        },
        body: JSON.stringify({ prompt}),
        }).then(res => res.json()).then((res: any) => {
            onGenerate(res.response);
        }).finally(() => {
            setLoading(false);
        })
    }

    return <div className="flex flex-col">
        <Input.TextArea style={{resize: 'none'}} className='resize-none' placeholder="please input your prompt to generate the cron expression" onChange={(e) => setPrompt(e.target.value)} value={prompt} rows={4} />
        <button className={`${loading ? "cursor-wait" : ""} bg-[length:200%_100%] hover:animate-ai-animate-btn px-4 py-1 flex justify-center items-center mt-4 mb-4 self-end bg-ai text-white rounded-md`  }onClick={btnClick}>
            <span className='mr-2'>Run</span>
            <Image alt='image' width={10} height={10} src={"/magic.svg"} />
        </button>
    </div>
}

5. AI Integration

5.1 API Route Setup

The cron API route (referenced from app/api/cron/route.ts) handles:

export async function POST(request: NextRequest) {
  // let responseText = 'Hello World'

  const { prompt } = await request.json<any>();
  const input = { prompt: `
      role: you are a cron expression master;
      capability: you can give valid cron expression by user nature language input;
      examples:
      Q: every minute invoke cron expression
      A: * * * * *

      Q: every hour at 30 minutes cron expression
      A: 30 * * * *

      Q: once a day cron expression
      A: 0 0 * * *

      Q: once a month
      A: 0 0 1 * *

      limit:
      1. no "?" or "#"  non-standard character.
      2. all cron expression only have 5 part(eg * * * * *).

      rule: 
      1. don't give me expression detail just give me result.
      2. the result must be a valid cron expression. 
      3. no extra useless string, just a cron expression (very important).

      now please generate cron by user input "${prompt}", if you can't give valid cron expression just told me "please try again"
    `};

  //@ts-ignore
  const answer = await getRequestContext().env.AI.run(
    "@cf/meta/llama-3-8b-instruct",
    input,
  );

  return new Response(JSON.stringify(answer))
}

5.2 AI Prompt Template

The AI system uses a structured prompt template that includes:

Role definition
Capabilities
Examples
Rules and limitations

Example prompt structure:

const input = {
  prompt: `
    role: you are a cron expression master;
    capability: you can give valid cron expression by user nature language input;
    examples: [...]
    limit: [...]
    rule: [...]
  `
}

5.3 Error Handling

Implement error handling for:

Invalid AI responses
Network failures
Invalid cron expressions

6. Deployment

6.1 Deployment Steps

Push code to GitHub repository
Connect repository to Cloudflare Pages
Configure build settings:

   Build command: npm run build
   Build output directory: .next

6.2 Environment Variables

Required environment variables:

AI: Cloudflare AI binding
NODE_VERSION: 18.x or higher

Best Practices

Use edge runtime for optimal performance
Implement proper error handling
Add input validation
Use TypeScript for type safety
Follow responsive design principles
Implement proper loading states

Limitations

AI response time may vary
Limited to standard cron expressions (5 parts)
No support for special characters (?, #)
Edge runtime restrictions

Conclusion

This project demonstrates how to build a modern AI-powered web application using Cloudflare's ecosystem and Next.js. The combination provides a scalable, performant solution for generating cron expressions through both visual interface and AI assistance.

For more information, refer to:

Top comments (1)

WayneTyler • Nov 28

This is a cool project! I love how Cloudflare Pages and Next.js come together to make something like this so powerful. If you’re looking for more hosting options, Cloudways could be a great fit for those looking to run similar Next.js applications on a managed platform with flexibility. DigitalOcean or Vercel are also solid options for quick deployments.

DEV Community