anysearch-ai/anysearch-skill

What it solves

AnySearch Skill provides a unified real-time search engine interface for AI agents, allowing them to access the web, vertical domain searches, and full-page content extraction. It eliminates the need for agents to manually handle multiple search APIs or scraping tools by providing a single, multi-runtime CLI toolset.

How it works

The project provides a set of CLI scripts (available in Python, Node.js, Bash, and PowerShell) that agents can execute to perform searches. The agent detects the available runtime on the system and persists the preferred command to a runtime.conf file. Once configured, the agent can call specific subcommands like search, batch_search, and extract to retrieve information from the web in JSON or Markdown format.

Who it’s for

This tool is designed for AI agent developers and platforms (such as Claude Code, Cursor, Windsurf, and OpenCode) that allow agents to install and use external "skills" or tools to extend their capabilities.

Highlights

• Multi-Runtime Support: Works across Python, Node.js, Shell, and PowerShell to ensure compatibility across different environments.
• Automated Registration: Agents can register users and obtain API keys automatically via a single API call using only an email address.
• Full-Page Extraction: Includes a dedicated extract command to convert web pages into Markdown for easier agent consumption.
• Parallel Batch Search: Supports executing multiple search queries simultaneously to increase efficiency.
• Flexible Authentication: Offers both anonymous access (with lower limits) and API key-based access for higher quotas.