AI Meme Search Engine

A meme search engine powered by semantic search technology. This application allows you to search for memes using natural language, finding both direct text matches and semantically similar content.

memesearcher.mov

Features

• Semantic Search: Find memes based on their semantic meaning, not just keywords
• Direct Matching: Standard text-based search is also supported
• Fast Performance: Results are cached for quick repeat searches
• Rate Limiting: Protects the API from abuse

Tech Stack

• Next.js 15: For the frontend and API routes
• Vercel Blob: For meme image storage
• Upstash Vector: For vector embeddings storage
• Upstash Redis: For caching and rate limiting
• OpenAI Embeddings: For generating semantic embeddings

Implementation Details

This project is based on Vercel Labs' Semantic Image Search, but with several key differences:

1. Uses Upstash Vector instead of PostgreSQL/pgvector for storing embeddings
2. Uses Upstash Redis for caching search results and implementing rate limiting
3. Optimized for meme content rather than general images

Getting Started

Prerequisites

You'll need accounts with:

• Vercel
• Upstash (for both Vector and Redis)
• OpenAI (for embeddings)

Environment Setup

Create a .env.local file with the following variables:

# OpenAI
OPENAI_API_KEY=

# Vercel Blob Storage
BLOB_READ_WRITE_TOKEN=

# Upstash Vector
UPSTASH_VECTOR_URL=
UPSTASH_VECTOR_TOKEN=

# Upstash Redis
UPSTASH_REDIS_REST_URL=
UPSTASH_REDIS_REST_TOKEN=

# Redis Cache TTL in seconds (24 hours)
REDIS_CACHE_TTL=86400

# Rate Limiting
RATE_LIMIT_REQUESTS=50
RATE_LIMIT_WINDOW=60

Installation

npm install
npm run dev

Deployment

Deploy to Vercel with:

vercel

Be sure to set up all the environment variables in your Vercel project settings.

How It Works

1. User enters a search query
2. System checks Redis cache for existing results
3. If no cache hit, the system:
   • Generates embeddings from the query
   • Searches for direct matches in titles and descriptions
   • Performs vector similarity search for semantic matches
   • Combines and deduplicates results
   • Caches the results in Redis
4. Results are displayed to the user