All posts
May 18, 2026

Build Smarter Apps with Union-Search-Skill: 20+ Engines in One

Streamline your AI development with union-search-skill. Access GitHub, Reddit, TikTok, Bilibili, and 20+ platforms via a unified search interface.

The Problem with Fragmented Search

Modern applications often need to pull data from diverse sources: developers want GitHub issues, researchers need Reddit discussions, and marketers track trends on platforms like TikTok or Xiaohongshu. Traditionally, this meant managing twenty different APIs, twenty different authentication flows, and twenty different data schemas.

This is where union-search-skill changes the game. It is a modular AI building block designed to act as a universal bridge across the fragmented web. By providing a standardized interface for search, it allows builders to focus on what to do with the data rather than how to fetch it.

What is Union-Search-Skill?

Developed by author runningZ1, the union-search-skill is a comprehensive search aggregator. It isn't just a wrapper for Google; it is a deep integration into niche and high-value platforms across the Western and Chinese internet ecosystems.

Whether you are building a competitive intelligence tool, a sentiment analysis bot, or an automated researcher, this skill provides structured Markdown or JSON outputs that are perfectly suited for LLM ingestion.

Key Supported Platforms:

  • Developer Hubs: GitHub (Repos, Code, Issues) and Reddit.
  • Social Media: Twitter (X), YouTube, Bilibili, Douyin (TikTok China), and Xiaohongshu.
  • AI-Native Search: Tavily, Metaso, and Exa (Neural Search).
  • No-API Alternatives: DuckDuckGo, Brave, and Wikipedia for cost-effective scraping.
  • Content & Images: WeChat articles and batch image downloading from 18 platforms including Unsplash and Baidu.

How to Install

Integrating the union-search-skill into your project is straightforward via the Lovable CLI.

lovable add union-search-skill

After installation, you will need to configure your credentials. The skill provides an ENV_TEMPLATE.txt that you can copy to a .env file to manage your various API keys for premium services like Google Custom Search or Tavily.

Use Cases for Builders

1. AI-Driven Market Research

Use the --group social flag to simultaneously scan Weibo, Xiaohongshu, and Twitter for mention of a brand. The skill's ability to normalize these disparate sources into a single response makes it easy to pipe the results into a GPT-4 summarizer.

2. Technical Documentation Assistants

By combining GitHub and StackOverflow (via general search), you can build a coding assistant that has the latest context from both source code and community discussions.

3. Automated Image Dataset Creation

The included union_image_search module allows for bulk downloading with metadata preservation—ideal for fine-tuning computer vision models.

Example Usage via CLI

Once configured, you can perform complex multi-platform queries from your terminal or call them from your application logic:

# Search across developer platforms for AI Agent trends
python union_search_cli.py search "AI Agent" --group dev --preset large

# Targeted search for a specific platform with a limit
python union_search_cli.py platform github "deep learning library" --limit 5

# Extract clean Markdown content from a URL
python union_search_cli.py defuddle https://lovable.dev/skills --json

Technical Advantages

What sets this skill apart is its standardization. Every module follows a strict input/output convention. You don't have to worry about one API returning results in a nested list and another in an object; union-search-skill handles the transformation.

Furthermore, the "Doctor" command serves as a built-in health check for your integration:

python union_search_cli.py doctor

This ensures all your API keys are valid and your network dependencies are met before you deploy to production.

Pro-Tips for Optimization

  • Use Presets: Instead of manually defining limits, use --preset small (3 results) or --preset extra (20 results) to quickly balance speed and depth.
  • Deduplication: When searching multiple engines, use the --deduplicate flag to ensure your LLM isn't processing the same URL twice, saving you tokens.
  • Structured Output: Always use the --pretty or --json flags if you are piping the output into another script to ensure parseability.

Bring the entire web's knowledge into your app with one modular skill.

Explore more and get started at /skill/union-search-skill

Try the skill
View union-search-skill →

Related posts