Tavily MCP Server

A powerful Model Context Protocol (MCP) server that brings Tavily's AI-powered search and content extraction capabilities directly to your AI applications. This server enables intelligent web searching with advanced filtering options and clean content extraction from any URL, making it easy to retrieve and process web information for AI workflows.

Overview

The Tavily MCP server provides seamless integration with Tavily's search and extraction APIs, allowing AI assistants to perform sophisticated web searches and extract clean, structured content from web pages. Whether you need real-time news updates, comprehensive research across the web, or parsed content from specific URLs, this server delivers high-quality results optimized for AI consumption.

Features

AI-Powered Search

Leverage Tavily's intelligent search engine designed specifically for AI applications. The search functionality goes beyond traditional web search by understanding context and intent, delivering results that are pre-processed and optimized for language models.

Advanced Content Extraction

Extract and parse content from any URL with automatic cleaning and formatting. The extraction engine removes clutter, advertisements, and irrelevant elements, returning only the meaningful content in your preferred format.

Flexible Filtering Options

Control your search results with precision through comprehensive filtering capabilities including date ranges, domain inclusion/exclusion, country-specific boosting, and topic categorization.

Multiple Search Topics

Choose from specialized search modes including general web search, real-time news monitoring, and finance-specific searches to get the most relevant results for your use case.

Rich Result Customization

Tailor the search output to your needs with options to include AI-generated answers, images with descriptions, favicons, raw HTML content, and configurable result depths.

Tools

tavily_search

Perform an AI-powered web search using Tavily with comprehensive customization options.

Parameters:

• query (required, string): The search query to execute
• topic (string): Category of search - general for broad searches, news for real-time updates, finance for financial data
• search_depth (string): Choose between basic (1 credit) or advanced (2 credits) for more relevant and comprehensive results
• max_results (integer, 0-20): Maximum number of search results to return
• include_answer (boolean): Include an LLM-generated answer to your query
• include_images (boolean): Include relevant image search results
• include_image_descriptions (boolean): Add AI-generated descriptions to images (requires include_images)
• include_favicon (boolean): Include favicon URLs for each result
• include_raw_content (boolean): Include cleaned HTML content of search results
• time_range (string): Filter results by publication date - day, week, month, or year (also accepts d, w, m, y)
• start_date (string): Start date filter in YYYY-MM-DD format
• end_date (string): End date filter in YYYY-MM-DD format
• days (integer, minimum 1): Number of days back to search (news topic only)
• country (string): Boost results from a specific country (general topic only) - accepts any country name from Afghanistan to Zimbabwe
• include_domains (array of strings, max 300): List of domains to specifically include in results
• exclude_domains (array of strings, max 150): List of domains to exclude from results
• chunks_per_source (integer, 1-3): Number of content chunks per source (advanced search only)
• auto_parameters (boolean): Enable automatic parameter configuration based on query intent

tavily_extract

Extract and parse content from one or more URLs, returning clean, formatted content ready for AI processing.

Parameters:

• urls (required, array of strings): List of URLs to extract content from (must be at least one valid URI)
• format (string): Output format - markdown or text
• extract_depth (string): Choose between basic (1 credit per 5 URLs) or advanced (2 credits per 5 URLs) for enhanced extraction including tables and embedded content
• include_images (boolean): Include a list of images extracted from the URLs
• include_favicon (boolean): Include favicon URLs for the extracted pages
• timeout (number, 1-60): Maximum wait time in seconds for the extraction to complete

Resource Templates

search

URI Template: tavily://search/{query}

Access and retrieve search results for a specific query through the resource system. This template allows you to reference search results as resources within your AI application, enabling consistent access to search data across different contexts.

extracted-content

URI Template: tavily://extracted-content/{url}

Access extracted and parsed content from a specific URL as a resource. This template provides a standardized way to reference cleaned web content, making it easy to work with extracted data in your AI workflows.

Use Cases

Research and Information Gathering

Quickly gather comprehensive information on any topic with AI-optimized search results. The advanced search depth ensures you receive the most relevant and authoritative sources.

Real-Time News Monitoring

Stay updated with the latest news by using the news topic search with time-based filtering. Perfect for monitoring breaking stories or tracking developments over specific time periods.

Content Analysis

Extract clean content from web pages for analysis, summarization, or further processing. The extraction tool handles complex page layouts and returns only the meaningful content.

Domain-Specific Research

Use domain filtering to focus your research on specific authoritative sources or exclude unreliable domains, ensuring higher quality results.

Multi-Source Content Aggregation

Extract and compare content from multiple URLs simultaneously, making it easy to gather information from various sources in a single operation.

Financial Research

Leverage the finance topic search to get specialized results for market research, company information, and financial news.

Credit Usage

The Tavily MCP server uses a credit-based system:

• Basic Search: 1 credit per search
• Advanced Search: 2 credits per search (provides more relevant results and additional content chunks)
• Basic Extraction: 1 credit per 5 URLs
• Advanced Extraction: 2 credits per 5 URLs (includes tables and embedded content)

Choose the appropriate depth based on your quality requirements and credit availability.