tavily-search

The Tavily Search skill acts as an intelligent bridge between your AI agent and the live internet. By leveraging the Tavily API, specifically engineered for LLMs and agentic workflows, this skill enables your agent to perform context-aware research, verify facts, and stay current with fast-moving topics. It is designed for developers, researchers, and power users who require reliable, noise-free web results without the traditional overhead of standard search engines. Whether you are analyzing current technology trends, tracking global news, or gathering data for complex reports, this tool integrates seamlessly into your OpenClaw or MCP-compatible environments.

• Advanced search depth control: Switch between basic queries and deep-dive research modes to balance speed with information granularity.

• Direct content extraction: Retrieves clean, parseable text directly from web pages, bypassing standard boilerplate content and advertisements.

• AI-ready output: Returns structured data including search answers, source citations, and optional raw content for deeper analysis.

• Flexible configuration: Supports both environment variables and standardized JSON configuration files, making it easy to manage across different deployments.

• Multi-faceted research: Handles news, technical updates, and thematic investigation requests efficiently.

• Requires a valid Tavily API key; obtain one from tavily.com to start usage.

• Best suited for search-heavy tasks; avoid using for static URL fetching where a specialized web_fetch tool is available.

• Highly customizable query parameters: Supports setting result limits (up to 10), enabling AI-generated summaries, and toggling image retrieval.

• Useful for verifying real-time information: Great for avoiding the knowledge cutoff issues inherent in static large language models.

• Error handling: Always ensure your TAVILY_API_KEY environment variable or openclaw.json configuration is correctly initialized to prevent connectivity issues.