How do the Mini and Pro models differ?

The Mini model is designed for fast, targeted questions, while the Pro model is optimized for comprehensive, multi-angle analysis of complex topics.

Does this skill require a separate API key?

Yes, you need a valid Tavily API key, which must be added to your Claude Code settings to enable web-based research.

Can I export research results to a specific file format?

Yes, you can use the --output flag to save your research results directly to a file, which is especially useful for JSON-structured reports.

Can I use this skill for automated data pipelines?

Absolutely. By providing a JSON schema, you can ensure the research output is structured and predictable, making it easy to parse for downstream applications.

AI Web Research

Name: AI Web Research
Author: tavily-ai

bytavily-ai

•

Web Scraping & Data Collection

Conducts comprehensive, web-grounded research with automatic citations and structured data output directly from your terminal.

About

The Research skill empowers Claude Code to perform deep-dive investigations by leveraging the Tavily API to gather, synthesize, and cite real-time web data. It is an essential tool for developers and analysts who need to generate technical comparisons, market reports, or grounded documentation without leaving their command-line environment. By supporting custom JSON schemas and background execution, it allows users to transform complex web searches into structured, actionable insights ready for integration into larger workflows or pipelines.

Key Features

Flexible model selection with Mini for speed and Pro for multi-angle depth
Support for structured JSON output via custom schemas for predictable data
Streaming mode and background execution for uninterrupted terminal workflows
Multiple citation formats including MLA, APA, Chicago, and numbered
1 GitHub stars
AI-synthesized research with automatic web source gathering and analysis

Use Cases

Building data pipelines that require real-time, cited AI insights in a structured format
Conducting automated market analysis and startup landscape research
Generating detailed technical comparisons between different software frameworks or libraries

About

Key Features

Flexible model selection with Mini for speed and Pro for multi-angle depth
Support for structured JSON output via custom schemas for predictable data
Streaming mode and background execution for uninterrupted terminal workflows
Multiple citation formats including MLA, APA, Chicago, and numbered
1 GitHub stars
AI-synthesized research with automatic web source gathering and analysis

Use Cases

Building data pipelines that require real-time, cited AI insights in a structured format
Conducting automated market analysis and startup landscape research
Generating detailed technical comparisons between different software frameworks or libraries