cloudinary--windsurf-visual-search-demo

An application demonstrating how to build a visual media search UI. Users can drag and drop an image to trigger Cloudinary's MCP servers, which use AI to extract features, query for visually similar assets, and return the results, streamlining the documentation workflow for developers.

GitHubスター

0

ユーザー評価

未評価

お気に入り

0

閲覧数

4

フォーク

0

イシュー

0

README
Smart Documentation Assistant (Visual Search Demo)

This project is a self-contained demo that simulates a Smart Documentation Assistant inside a developer's code editor (like Windsurf). It demonstrates how developers can use AI-powered visual search to find media assets and insert them into documentation without leaving the editor.


🧠 Core Workflow
  1. A developer is writing a Markdown file.
  2. They drag and drop a local image (e.g., a rough sketch or screenshot) into a search panel.
  3. The system uses AI to find visually similar, official assets from a Cloudinary media library.
  4. The developer clicks a result, and the correctly formatted Markdown for that image is inserted into their document.

▶️ How to Run the Demo

This is a simple front-end demo — no installation or build steps required.

  1. Clone or download the project files.
  2. Open index.html directly in your web browser.
  3. You'll see the simulated editor. Drag any image from your computer onto the Drag & Drop zone to trigger the visual search.

🛠 Tech Stack (Simulated)
  • Editor Environment: Windsurf

  • AI & Media Backend: Cloudinary MCP Servers

    • @cloudinary/asset-management – for uploads and searching
    • @cloudinary/analysis – for AI-powered visual feature extraction
  • Frontend: HTML, CSS, Vanilla JavaScript

Note: The calls to the Cloudinary MCP servers are simulated in the app.js file. Results are mocked to showcase the intended workflow.