| name | tavily-search |
| description | Web search via Tavily API for research and fact-finding. |
Tavily Search
Web search, content extraction, and deep research via Tavily MCP.
When to Use
- Web search — real-time information, news, documentation lookups
- Content extraction — pull structured content from known URLs
- Replacement/complement for Exa when Exa returns 402 errors
Available Tools
| Tool | Purpose |
|---|
tavily-search | Web search with filtering, domain scoping, depth control |
tavily-extract | Extract readable content from one or more URLs |
tavily-search Parameters
| Param | Type | Default | Notes |
|---|
query | string | required | Search query |
search_depth | "basic" | "advanced" | "basic" | Advanced = deeper crawl, slower |
topic | "general" | "news" | "finance" | "general" | Topic category |
max_results | number | 5 | Max results (1-20) |
include_domains | string[] | — | Restrict to these domains |
exclude_domains | string[] | — | Exclude these domains |
include_raw_content | boolean | false | Include full page content |
include_images | boolean | false | Include image URLs |
days | number | — | Recency filter (last N days) |
tavily-extract Parameters
| Param | Type | Notes |
|---|
urls | string[] | URLs to extract content from |
Usage Patterns
Quick web search
tavily-search({ query: "Effect-TS Schema.TaggedStruct examples" })
News search with recency
tavily-search({ query: "AI agent frameworks", topic: "news", days: 7 })
Domain-scoped search
tavily-search({
query: "useEffect cleanup",
include_domains: ["react.dev", "github.com"],
search_depth: "advanced"
})
Extract content from URL
tavily-extract({ urls: ["https://effect.website/docs/schema/classes"] })
Search Tool Routing (updated)
| Need | Primary | Fallback |
|---|
| General web search | tavily-search | firecrawl_search, web_search |
| Code examples | exa_get_code_context | tavily-search with domain filter |
| Company research | exa_company_research | tavily-search with topic |
| Deep research | nia_research | tavily-search advanced depth |
| GitHub repo docs | deepwiki | nia_tracer |
| Library API docs | context7, effect-docs | tavily-extract on docs URL |
| Scrape specific page | firecrawl_scrape | tavily-extract |