JUHE API Marketplace

Build Auto-Enhanced Listings with AI Visual Pipeline for Second-Hand Platforms

3 min read
By Chloe Anderson

Introduction

Automating listing visuals for second-hand platforms can significantly improve item appeal, reduce manual editing, and increase conversion rates. For CTOs, product managers, and engineering teams, a robust AI pipeline delivers scalable, consistent outputs.

Why Auto-Enhancement Matters

Benefits for Second-Hand Platforms

  • Consistent quality across diverse product images
  • Improved buyer confidence through better visuals
  • Faster turnaround to publish listings

Cost Advantage

Nano Banana Pricing

  • Official standard model: $0.039/image
  • Our standard quality rate: $0.02/image, official-grade consistency within about 10 seconds
  • Official Pro model: $0.134/image
  • Our Pro rate: $0.068/image
  • For teams processing thousands of images daily, savings add up quickly

Sora AI Video Pricing

  • Official rate: $1–$1.5 per video
  • Our rate: $0.12 per video
  • Suitable for short promo clips integrated into listings

AI Visual Pipeline Components

Step 1: Image Enhancement with Nano Banana

The gemini-2.5-flash-image model offers fast processing for standard needs, while gemini-3-pro-image-preview suits premium outputs. Integrated with an LLM, descriptions can auto-update based on enhanced visuals.

Step 2: Video Creation with Sora

Sora’s sora-2 model generates short videos efficiently. Asynchronous API design allows progress checks without blocking main workflows.

Workflow Architecture

API Flow

  1. Input raw product image
  2. Process through Nano Banana model
  3. Return optimized base64 image data
  4. LLM creates enhanced copy
  5. Sora generates optional promo video

Scalability Features

  • Stable performance under high concurrency
  • Parallelization and microservice compatibility

Hands-On API Example

Image Generation Request

curl --location --request POST 'https://wisdom-gate.juheapi.com/v1/chat/completions' \
--header 'Authorization: YOUR_API_KEY' \
--header 'Content-Type: application/json' \
--header 'Accept: */*' \
--data-raw '{
  "model": "gemini-2.5-flash-image",
  "messages": [{"role": "user","content": [{"text": "generate a high-quality image.","type": "text"},
        {"image_url": {"url": "https://blog-images.juhedata.cloud/example.jpeg"},"type": "image_url/base64"}]}
  ],
  "stream": false
}'

Video Creation Request

curl -X POST "https://wisdom-gate.juheapi.com/v1/videos" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: multipart/form-data" \
  -F model="sora-2" \
  -F prompt="A serene lake surrounded by mountains at sunset" \
  -F seconds="15"

Video Progress Check

curl -X GET "https://wisdom-gate.juheapi.com/v1/videos/{task_id}" \
  -H "Authorization: Bearer YOUR_API_KEY"

Integration Tips

Optimize Request Batching

  • Group multiple images in requests for efficiency
  • Use caching to avoid reprocessing identical inputs

Monitor and Alerts

  • Track API latency and error rates
  • Set up retries for intermittent failures

Scaling Up

Horizontal Scaling

  • Deploy multiple processing nodes
  • Use load balancers for even traffic distribution

Data Flow

  • Employ message queues like RabbitMQ or Kafka
  • Implement failover handling for robust job completion

Security Considerations

  • Store API keys securely
  • Enforce per-team rate limits

Performance Benchmarks

Image Pipeline

  • Average 10 seconds for base64 output
  • Stable under 10,000 concurrent requests

Video Pipeline

  • ~20 seconds for a 15-second clip
  • Resilient under high load

Measuring Impact

KPIs

  • CTR increase from enhanced visuals
  • Reduction in listing prep time
  • Cost savings per 1,000 items processed

Conclusion

Combining Nano Banana for image refinement with Sora for compelling video creates a complete AI-powered visual pipeline. By leveraging reduced pricing and stable performance, teams can scale content creation without sacrificing quality or speed.