search

Search web content via Tavily's LLM-optimized API with configurable parameters.

Updated Feb 2, 2026

Search Tags:#API #search #LLM #metadata #web #prompting #snippets

Ready to run

npx skills add https://github.com/drillan/marketschema --skill search-drillan

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: search
Source: https://github.com/drillan/marketschema/tree/main/.claude/skills/search
Command: npx skills add https://github.com/drillan/marketschema --skill search-drillan

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires jq, curl, and includes scripts (resource) components.

What problem does it solve?

This Skill enables fast retrieval of relevant web content via Tavily's LLM-optimized search API, returning snippets, scores, and metadata so you can answer questions without writing code.

Core Features & Use Cases

LLM-optimized search: Returns concise, snippet-rich results with scores suitable for prompt-based workflows.
Metadata-rich results: Includes URLs, domains, and scoring metadata to support verification and downstream processing.
Use Case: Quickly gather background information or verify facts for AI-assisted responses.

Quick Start

Use the search script to fetch the top results for your query and integrate the snippets into your assistant's context for summarization or decision making.

Frequently Asked Questions about search

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I get LLM-optimized web search results for prompt-based workflows?▼

LLM-optimized web search results are retrieved by sending a query via Tavily's API, returning concise snippets, metadata, and domain scores ready for prompt-based workflows. The script formats the JSON payload to fetch background briefs and fact-checking summaries.

What is the best way to include raw web content in an AI-assisted research task?▼

Raw web content is included by setting the include_raw_content parameter in the search payload. The script fetches the requested URLs and returns the full content alongside snippet-rich summaries to support evidence-based AI responses.

Can I filter web search results by specific domains and time ranges?▼

Web search results can be filtered by specific domains and time ranges using the include_domains and time_range parameters. The script passes these filters in the JSON payload to Tavily's API to restrict the returned metadata and snippets.

Do I need curl and jq to run web search scripts?▼

You need curl and jq installed to execute the search script. Curl handles the HTTP request with the Authorization header, while jq parses the returned JSON metadata, URLs, and scoring data from the API response.

How does metadata-rich search support fact-checking and verification?▼

Metadata-rich search supports fact-checking by returning URLs, domains, and scoring data alongside snippets. This metadata allows you to verify source relevance and process the retrieved web content for downstream evidence-based responses.

What are the limitations of using Tavily's API for snippet retrieval?▼

Using Tavily's API for snippet retrieval requires an active API key configured in the Authorization header. The search depth and max_results parameters constrain the volume of metadata and snippets returned per query.