I Built a Universal AI CLI Tool in 10 Minutes (Inspired by Composio's Product Hunt Launch)

#cli #python #javascript #ai

Composio's Universal CLI just launched on Product Hunt and it's blowing up. The concept is simple: connect your AI agents to 500+ tools from the command line. Developers love it.

But I thought — what if we take this further? What if your CLI could not just orchestrate tools, but generate images, videos, and audio — all from your terminal?

So I built it. In 10 minutes. Using NexaAPI.

What I Built

A Universal AI CLI that does:

# Generate an image
python ai_cli.py image "a futuristic city at sunset"
# ✅ Image saved to output.png — Cost: $0.003

# Convert text to speech
python ai_cli.py audio "Hello, this is your AI assistant" --voice nova
# ✅ Audio saved to output.mp3

# Generate a short video
python ai_cli.py video "a cat playing piano in a jazz club"
# ✅ Video saved to output.mp4

Why NexaAPI?

I tried a few options. Here's what I found:

API	Image Cost	Models	Single Key?
NexaAPI	$0.003	50+	✅
OpenAI DALL-E 3	$0.04	2	✅
Replicate	$0.003+	100+	✅
fal.ai	$0.003+	50+	✅

NexaAPI wins on simplicity: one API key for images, video, AND audio. Available on RapidAPI — subscribe in seconds.

The Code (Python)

# pip install nexaapi
import argparse
from nexaapi import NexaAPI

client = NexaAPI(api_key='YOUR_RAPIDAPI_KEY')

def generate_image(prompt, output='output.png', model='flux-schnell'):
    result = client.image.generate(model=model, prompt=prompt)
    with open(output, 'wb') as f:
        f.write(result.image_data)
    print(f'✅ Image saved to {output} — Cost: $0.003')

def generate_audio(text, output='output.mp3', voice='alloy'):
    result = client.audio.tts(text=text, voice=voice)
    with open(output, 'wb') as f:
        f.write(result.audio_data)
    print(f'✅ Audio saved to {output}')

def generate_video(prompt, output='output.mp4'):
    result = client.video.generate(model='kling-v1', prompt=prompt)
    with open(output, 'wb') as f:
        f.write(result.video_data)
    print(f'✅ Video saved to {output}')

def main():
    parser = argparse.ArgumentParser(description='Universal AI CLI — https://nexa-api.com')
    subparsers = parser.add_subparsers(dest='command')

    img = subparsers.add_parser('image')
    img.add_argument('prompt')
    img.add_argument('--output', default='output.png')
    img.add_argument('--model', default='flux-schnell')

    audio = subparsers.add_parser('audio')
    audio.add_argument('text')
    audio.add_argument('--output', default='output.mp3')
    audio.add_argument('--voice', default='alloy')

    video = subparsers.add_parser('video')
    video.add_argument('prompt')
    video.add_argument('--output', default='output.mp4')

    args = parser.parse_args()
    if args.command == 'image':
        generate_image(args.prompt, args.output, args.model)
    elif args.command == 'audio':
        generate_audio(args.text, args.output, args.voice)
    elif args.command == 'video':
        generate_video(args.prompt, args.output)

if __name__ == '__main__':
    main()

The Code (JavaScript/Node.js)

// npm install nexaapi commander
import NexaAPI from 'nexaapi';
import { writeFileSync } from 'fs';
import { program } from 'commander';

const client = new NexaAPI({ apiKey: process.env.NEXAAPI_KEY });

program.command('image <prompt>')
  .option('-o, --output <file>', 'Output file', 'output.png')
  .option('-m, --model <model>', 'Model', 'flux-schnell')
  .action(async (prompt, opts) => {
    const result = await client.image.generate({ model: opts.model, prompt });
    writeFileSync(opts.output, result.imageData);
    console.log(`✅ Image saved to ${opts.output} — Cost: $0.003`);
  });

program.command('audio <text>')
  .option('-o, --output <file>', 'Output file', 'output.mp3')
  .option('-v, --voice <voice>', 'Voice', 'alloy')
  .action(async (text, opts) => {
    const result = await client.audio.tts({ text, voice: opts.voice });
    writeFileSync(opts.output, result.audioData);
    console.log(`✅ Audio saved to ${opts.output}`);
  });

program.command('video <prompt>')
  .option('-o, --output <file>', 'Output file', 'output.mp4')
  .action(async (prompt, opts) => {
    const result = await client.video.generate({ model: 'kling-v1', prompt });
    writeFileSync(opts.output, result.videoData);
    console.log(`✅ Video saved to ${opts.output}`);
  });

program.parse();

Real-World Use Cases I've Found

1. CI/CD thumbnails — auto-generate blog post thumbnails on every publish
2. Batch product images — generate 100 product variations overnight
3. Podcast production — convert blog posts to audio automatically
4. Game asset generation — generate sprites and backgrounds from prompts