Pickify

Main
Recommend Combination
Chromewebstore

Pickify is a project designed to address the challenges consumers face while shopping online.

Inspiration

Traditional image search methods require users to switch to Google Lens via Google Home, which can be inconvenient. Additionally, finding desired products, exploring related items, and completing purchases often take significant time and effort.

Pickify streamlines this process, enabling consumers to discover and purchase products more intuitively and efficiently.

What it does

Similar Product Recommendation:
Users can capture and upload an image of a product to receive recommendations for similar items. Clicking on a recommended product redirects users to the corresponding product page.
Custom Combination Recommendation:
Analyzes the images users have captured to recommend related products that would complement the selected items.

How we built it

Pickify was developed as a Chrome extension using the Gemini Build-in API for advanced image analysis and recommendation algorithms.

Users capture and upload an image to GCP Cloud Storage.
The Built-in API - Prompt API analyzes the image and extracts keywords.
Using the extracted keywords, Gemini Flash 1.5 generates search terms and retrieves image URLs.
The keywords are mapped to higher-level categories based on the Google Product Taxonomy.
The extracted keywords, URLs, and categories are stored in a MongoDB database.
The Built-in API - Prompt API analyzes the current image along with previously captured data to generate combination keywords.
Recommendations are displayed in two tabs: one for similar products and another for related combinations.
The backend infrastructure is built using Google Cloud Platform (GCP).

Sequence Diagram

1. Capture Search

2. Recommend Combination

Challenges we ran into

Integrating the Gemini API: Adjusting the API to deliver accurate and meaningful recommendations was a challenging process.
User Experience Design: Ensuring the extension is easy to use while maintaining a smooth browsing experience.
Backend Optimization: Structuring data flows and improving response times for real-time product recommendations.

Accomplishments that we're proud of

Successfully implemented real-time image analysis and recommendation features.
Designed an intuitive UI that enhances user engagement.
Provided a practical tool to bridge the gap between browsing and purchasing decisions.

What we learned

Using Built-in APIs: Learned how to effectively apply Built-in APIs to real-world applications.
User-Centered Development: Gained insight into designing services that prioritize user needs and behaviors.
Technical Optimization: Improved performance to deliver a fast and accurate user experience.

What's next for Pickify

Improving Response Speed: Further optimizing the extension to provide faster recommendations.
Global Expansion: Supporting multiple languages using the Built-in API - Translator API to enhance accessibility for a global audience.
More accuracy: Improved precision of results, avoiding irrelevant shopping links and instead finding related products for purchase.