metacsv

• 

  Homepage

• 

  Homepage-2

• 

  Main Feature

• 

  Main Feature-2

Inspiration

As my friends are content creators, they experienced firsthand how time-consuming it is to manually generate metadata (titles, keywords, and descriptions) for each stock photo upload. Especially when working across multiple platforms—each with different CSV formats—it becomes tedious and error-prone. We built METACSV to automate this process using AI, allowing creators to focus more on content creation than on repetitive data entry.

What it does

METACSV allows users to upload images and automatically generate high-quality metadata using AI. With support for both default and customizable prompts, users can tailor the analysis to their content style. The tool then exports the metadata into CSV files formatted specifically for platforms like Adobe Stock, Shutterstock, Freepik, Canva, and Miricanvas. It also supports batch processing and user authentication for a seamless experience.

How we built it

Frontend: Vue.js + TailwindCSS Backend & Auth: Supabase (for authentication and API integration) AI Integration: Gemini API for image understanding and metadata generation Export Logic: Platform-specific CSV export functionality (Adobe, Freepik, etc.) Storage: Images are not stored; they are processed in-memory and sent directly to the API for analysis Batch Handling: Already implemented to process multiple images in one session

Challenges we ran into

Designing flexible CSV exports to meet the format requirements of multiple stock platforms Prompt tuning for consistent metadata output that aligns with SEO and platform guidelines Managing batch requests efficiently while preventing API rate-limiting or timeouts

Accomplishments that we're proud of

Built a functional end-to-end workflow: image upload → AI analysis → metadata export Integrated batch processing and user authentication from day one Achieved metadata formatting compatibility across multiple platforms—saving creators significant time

What we learned

Supabase is a fast and effective backend-as-a-service for rapid prototyping Prompt engineering plays a critical role in output quality for visual AI use cases Creator tools must be both powerful and simple to provide real value Using Bolt.new make app building take less time, so it more fast and efficient

What's next for metacsv

Add delayed request throttling and retry logic for batch API calls Integrate alternative AI models for platform-specific tuning (e.g., using a different model for Freepik vs. Shutterstock) Implement template-based metadata editing before export Explore direct platform submission APIs for a true end-to-end pipeline Add user-created prompt presets for different content niches

Built With

• bolt.new
• gemini
• netlify
• pinia
• supabase
• tailwindcss
• vite
• vue