Apify Store Searcher

A web application to search and browse the Apify Store with ML-enhanced category filtering.

Try It

Live URL: https://jirimoravcik--apify-store-searcher.apify.actor

Just open the link and start searching for Actors by name, category, or description.

Features

• Full-text search across all Actors in the Apify Store
• ML-powered category ranking using TF-IDF and cosine similarity to surface the most relevant Actors per category
• Smart agent detection - distinguishes true AI agents from scrapers that are miscategorized
• Neobrutalist UI with Apify's brand colors
• Real-time search with instant results
• Standby mode for always-on deployment on Apify platform

How It Works

ML Classification

The Actor fetches all Actors from the Apify Store API and computes ML relevance scores for each category using:

1. TF-IDF Vectorization - Converts Actor titles, descriptions, and metadata into numerical vectors
2. Exemplar Matching - Compares each Actor against hand-crafted exemplar descriptions for each category
3. Cosine Similarity - Ranks Actors by how closely they match the category exemplars
4. Metadata Signals - Applies adjustments based on:
   • Keywords in title/description (e.g., "agent", "scraper", "LLM")
   • Agentic payments whitelist status
   • MCP/Model Context Protocol mentions
   • Quality signals (user count, ratings, reviews)

This approach is particularly effective at filtering the Agents category, where many scrapers are miscategorized. True AI agents (with LLM integration, autonomous behavior, tool calling) are ranked higher than simple web scrapers.

Architecture

┌─────────────────┐     ┌──────────────────┐     ┌─────────────────┐
│   Frontend      │────▶│   HTTP Server    │────▶│   Store Cache   │
│   (HTML/JS)     │     │   (Python)       │     │   (In-memory)   │
└─────────────────┘     └──────────────────┘     └────────┬────────┘
                                                          │
                                                          ▼
                                                 ┌─────────────────┐
                                                 │  ML Classifier  │
                                                 │  (TF-IDF)       │
                                                 └─────────────────┘

• Frontend: Neobrutalist single-page app with search and category filters
• HTTP Server: Handles API requests and serves the frontend
• Store Cache: Fetches all Actors from API, caches with auto-refresh every 5 minutes
• ML Classifier: Computes category relevance scores using TF-IDF similarity

API Endpoints

Endpoint	Description
GET /	Serves the web frontend
GET /api/search	Search Actors with optional filters
GET /api/cache-stats	Get cache statistics
GET /health	Health check endpoint

Search Parameters

• q - Search query (searches title, description, username)
• category - Filter by category (e.g., AGENTS, AI, SOCIAL_MEDIA)
• limit - Max results to return (default: 50, max: 100)
• offset - Pagination offset (default: 0)

Categories

The following categories are supported:

Category	Description
AGENTS	AI agents with LLM integration and autonomous capabilities
AI	Machine learning and AI-powered tools
SOCIAL_MEDIA	Social media scrapers and tools
LEAD_GENERATION	Business contact and lead finding tools
ECOMMERCE	E-commerce and product data tools
SEO_TOOLS	SEO analysis and monitoring
JOBS	Job listing scrapers
MCP_SERVERS	Model Context Protocol servers
NEWS	News and media scrapers
REAL_ESTATE	Property and real estate data
DEVELOPER_TOOLS	Developer utilities
TRAVEL	Travel and booking data
VIDEOS	Video platform tools
AUTOMATION	Workflow automation tools
INTEGRATIONS	Platform integration tools
OPEN_SOURCE	Open source tools
OTHER	Miscellaneous tools

Local Development

Prerequisites

• Python 3.10+
• Apify CLI

Setup

# Clone the repository
git clone <repository-url>
cd apify-store-searcher

# Create virtual environment
python -m venv .venv
source .venv/bin/activate  # On Windows: .venv\Scripts\activate

# Install dependencies
pip install -r requirements.txt

# Run locally
apify run

The server will start at http://localhost:8080.

Deployment

Deploy to Apify Platform

# Login to Apify (requires API token)
apify login

# Deploy the Actor
apify push

The Actor runs in standby mode, staying alive to serve requests with low latency.

Tech Stack

• Python 3.10+ - Runtime
• Apify SDK - Actor framework and storage
• scikit-learn - TF-IDF vectorization and cosine similarity
• NumPy - Numerical computations
• httpx - Async HTTP client

License

ISC