Tavily

Search

search

Read-onlyIdempotent

AI/LLM-optimized web search built for RAG: returns a synthesized natural-language answer plus a ranked list of sourced results (title, url, content snippet, relevance score). Prefer this over scraping a generic search engine when you need grounded, citable web context. Example: search({ query: "latest SpaceX Starship test result" })

Input Schema

TableJSON Schema

Name	Required	Description
`query`	Yes	The search query / question to research.
`topic`	No	Search topic: "general" (default) or "news" for recent news coverage.
`_apiKey`	No	Optional — your own Tavily API key for higher limits; omit to use the shared Pipeworx key.
`max_results`	No	Maximum number of results to return (default 5, max 20).
`search_depth`	No	Search depth: "basic" (fast, default) or "advanced" (deeper, more thorough).
`include_answer`	No	Whether to include a synthesized AI answer string (default true).

Output Schema

TableJSON Schema

Name	Required	Description
`answer`	No	Synthesized answer; may be empty when nothing matched
`results`	Yes
`response_time`	No

Schema Changelog

Changes observed during successful MCP inspections.

Changed1 schema field changedAug 9, 2026

changedOutput schema / (root)

Previous value: -nullNew value: +{
+  "description": "Tavily web search. `answer` is a synthesized summary of the results.",
+  "properties": {
+    "answer": {
+      "description": "Synthesized answer; may be empty when nothing matched",
+      "type": [
+        "string",
+        "null"
+      ]
+    },
+    "response_time": {
+      "type": [
+        "number",
+        "null"
+      ]
+    },
+    "results": {
+      "items": {
+        "properties": {
+          "content": {
+            "type": "string"
+          },
+          "score": {
+            "type": [
+              "number",
+              "null"
+            ]
+          },
+          "title": {
+            "type": "string"
+          },
+          "url": {
+            "type": "string"
+          }
+        },
+        "required": [
+          "title",
+          "url"
+        ],
+        "type": "object"
+      },
+      "type": "array"
+    }
+  },
+  "required": [
+    "results"
+  ],
+  "type": "object"
+}

Changed1 schema field changedJul 29, 2026

addedInput schema / examples

Added value: +[
+  {
+    "query": "latest SpaceX Starship test result"
+  },
+  {
+    "include_answer": true,
+    "max_results": 10,
+    "query": "OpenAI GPT-5 release date",
+    "search_depth": "advanced",
+    "topic": "news"
+  }
+]

First observedJun 4, 2026

TDQS

A4.2/5.0

Behavior4/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Annotations already declare readOnlyHint, openWorldHint, idempotentHint, and destructiveHint false, covering safety. The description adds meaningful behavioral context: that it returns a synthesized answer and a ranked list with relevance scores, which is beyond the schema. It doesn't discuss rate limits or errors, but this is acceptable given annotation coverage.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness5/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is two sentences plus an example, with no wasted words. It front-loads the purpose, then adds usage guidance and a concrete example, all of which earn their place.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

An output schema exists, so the description needn't fully explain return values, yet it still conveys the key output structure (synthesized answer plus ranked sources). Combined with the schema, the overall package is complete for a search tool of this complexity, though it could mention which sibling scenarios it is not suited for.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters3/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema description coverage is 100%, with each of the six parameters described, including enums for topic and search_depth. The description provides an example but no additional parameter semantics beyond what the schema already provides, so the baseline score of 3 applies.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description states a specific verb-resource pair ('AI/LLM-optimized web search') and clearly defines the output: a synthesized natural-language answer plus a ranked list of sourced results with title, URL, snippet, and relevance score. It explicitly distinguishes the tool from scraping a generic search engine, which aligns with its sibling context.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines4/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

The description provides explicit guidance to prefer this tool over scraping a generic search engine when grounded, citable web context is needed. It does not contrast with sibling tools like deep_research or search_within, but the alternative it names is actionable and relevant.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

All Tavily tools Inspect

Try in Browser

One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

Glama MCP Gateway

Add one secure layer between your agents and this server.