Skip to content

tavily-ai/tavily

v1.0.1

Search, extract, crawl, map, and research the web with Tavily MCP tools and bundled skills for research, enrichment, monitoring, and integrations.

GitHubOpen Website

Tavily Power for Kiro

Search the web, extract content from URLs, crawl websites, and conduct AI-powered research — all from within Kiro. This power gives Kiro AI agents access to Tavily's LLM-optimized APIs for real-time web data.

About Tavily

Tavily is a search engine and API designed specifically for AI agents and LLMs. It delivers real-time web data — search results, page content, full-site crawls, and multi-source research reports — in formats optimized for agent consumption, with relevance scoring and metadata built in.

Tavily powers the retrieval layer for thousands of AI applications, from RAG pipelines to autonomous research agents.

Features

  • Web Search: Find current information with relevance scoring, domain filtering, and time-range controls
  • Content Extraction: Pull clean markdown or text from any URL, including JavaScript-rendered pages
  • Site Crawling: Extract content from entire websites with configurable depth, breadth, and semantic instructions
  • URL Mapping: Discover a site's URL structure without extracting content — fast recon before targeted extraction
  • AI-Powered Research: Comprehensive multi-source research reports with citations, powered by Tavily's research models

Installation

Prerequisites

  1. Tavily Account: Sign up at tavily.com

Kiro Power Setup

  1. Open the Powers panel in Kiro
  2. Click "Add power from GitHub"
  3. Enter: https://github.com/tavily-ai/tavily-power
  4. On first use, your browser will open for OAuth login — tokens are cached automatically

Alternatively, get an API key from your Tavily dashboard and set TAVILY_API_KEY in your environment.

What Can You Do with Tavily and Kiro?

1 — Search the Web

Use prompts like:

  • Search for the latest developments in quantum computing
  • Find recent articles about Rust async patterns on arxiv.org and github.com
  • What happened in AI regulation this month?

2 — Extract Content from URLs

Use prompts like:

3 — Crawl Documentation Sites

Use prompts like:

  • Crawl the docs at https://docs.example.com and find all API references
  • Get all the content under /guides/ on this documentation site
  • Map the URL structure of this site so I can decide what to extract

4 — Research a Topic

Use prompts like:

  • Research the competitive landscape for AI code assistants
  • What are the top 5 competitors to Notion in the SMB market?
  • Investigate the pros and cons of server-side rendering vs. static generation

Resources