tavily-search

Performs real-time web searches via the Tavily API and returns titles, URLs, and summaries.

Updated Jul 20, 2026
One-click install
npx skills add https://github.com/lsongdev/skills --skill tavily-search-lsongdev
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: tavily-search
Source: https://github.com/lsongdev/skills/tree/main/tavily-search
Command: npx skills add https://github.com/lsongdev/skills --skill tavily-search-lsongdev

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires tavily-python, and includes scripts (resource) components.

What problem does it solve? When users ask about recent events, official documentation, or time-sensitive facts, language models alone may guess or rely on outdated knowledge. This Skill retrieves live web results through the Tavily Search API so answers are grounded in verifiable, citable sources. ## Core Features & Use Cases - Real-Time Web Search: Query the internet for the latest news, version updates, announcements, and technology developments. - Resource Location: Find official documentation, original papers, course pages, and tool websites with direct links. - Flexible Output: Get human-readable text summaries or structured JSON for further programmatic processing, with an optional API-generated answer. - Use Case: A user asks "What are the new features in Python 3.13?" The Skill runs a Tavily search, gathers current articles and official release notes, and the answer is composed with source URLs cited. ## Quick Start Ask the assistant to look up a current topic, for example: search the web for the latest AI news this week and include source links.

Frequently Asked Questions about tavily-search

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

FAQPage Schema
How do I perform a web search with the Tavily API in Python?▼

Install the tavily-python package, set the TAVILY_API_KEY environment variable, then run the search script with a query string. It returns titles, URLs, and content snippets, with options for result count and JSON output.

How do I get JSON output from a Tavily search?▼

Pass the --format json flag to the search script. The full API response is printed as formatted JSON, which is suitable for parsing and further processing by a model or another program.

What Python dependencies does Tavily search require?▼

It requires Python 3 and the tavily-python package, installable via pip. A valid TAVILY_API_KEY environment variable must also be set, otherwise the script exits with an error message.

Why does the Tavily search script fail with an API key error?▼

The script checks for the TAVILY_API_KEY environment variable before calling the API. If it is missing, the script exits immediately; set the variable in your shell environment and rerun the command.

When should I use web search instead of relying on model knowledge?▼

Use web search for time-sensitive topics such as recent news, version releases, policy changes, or official documentation links. Model knowledge may be outdated, while search results provide current, citable sources.