Skip to main content

FrootAI — AmpliFAI your AI Ecosystem Get Started

All servers

Tavily Web Search & Extraction

tavily-ai

Real-time web search, page extraction, site mapping, and crawling via the official Tavily MCP server.

Installs
2.1k
Namespace
tavily-ai
Version
>=0.2.0 <1.0.0
Last attach
reviewed1 months ago

Add to your mcp.json

{
  "mcpServers": {
    "tavily-ai": {
      "command": "npx",
      "args": [
        "-y",
        "tavily-mcp"
      ],
      "env": {
        "TAVILY_API_KEY": "<TAVILY_API_KEY>"
      }
    }
  }
}
Open in VS CodeDon't have it? Install FrootAI for VS Code

Attach in a FrootAI play (fai-manifest)

mcp_scope:
  attached:
    - tavily-ai

Authentication

api-key

sign up at app.tavily.com, then set TAVILY_API_KEY to your `tvly-…` key; the free tier works for basic search.

Required credential: TAVILY_API_KEY

Environment variables

VariableRequiredAuth modeDescription
TAVILY_API_KEYyesapi-keyTavily API key (`tvly-…`). Issue from app.tavily.com; required for the local server to authenticate. Pass via env, never inline in args (doctrine #6).
DEFAULT_PARAMETERSnoanyOptional JSON object of default `tavily-search` parameters (e.g. include_images, max_results, search_depth) applied to every request unless overridden per call.
TAVILY_HUMAN_IDnoanyOptional opaque end-user identifier forwarded as the `X-Human-Id` header for per-user analytics; hashed server-side (SHA-256). Prefer an internal ID over raw PII.

Sample tools

  • tavily-ai.tavily-search

    Real-time web search with configurable depth, result count, and optional images.

  • tavily-ai.tavily-extract

    Extract clean, structured content from one or more web-page URLs.

  • tavily-ai.tavily-map

    Build a structured map of a website's pages from a root URL.

  • tavily-ai.tavily-crawl

    Systematically crawl a website starting from a base URL.

Used in recipes

FAI cookbook recipes that attach tavily-ai.

Destructive-action audit

No destructive tools — audited read-only

Destructive-action audit (X2.18): no destructive tools — search/extract/map/crawl are read-only web retrieval.

Known limitations

  • Requires a Tavily API key (TAVILY_API_KEY); without it the local server cannot authenticate and every tool returns an auth error.
  • Search, extract, map, and crawl all consume Tavily API credits — high `max_results` or deep crawls draw down the account quota faster.
  • Results reflect live public-web content and Tavily's index freshness; extraction quality varies with page structure (JS-heavy or paywalled pages may return partial content).

Spec contributors

Maintainers who landed commits onmcp-specs/tavily-ai.json.

  • pspsbali

Freshness

When this spec was last reviewed by a maintainer and last machine-validated against the live server.

Spec last reviewed
1 months ago(2026-06-25)
Auto-validated
never