Skip to main content
Glama
r28ai

Web Research to Docs

by r28ai
OverviewSchemaSourceRelated ServersDiscussionsAdmin
Python
Remote

tavily_search

Search the web in real time for AI agents when sources are unknown or current context is needed.

Instructions

Execute a real-time web search optimized for AI agents. Use when sources are unknown or current web context is needed. Prefer search_depth advanced with chunks_per_source 3 for stronger evidence per source.

Input Schema

TableJSON Schema
NameRequiredDescriptionDefault
queryYesSearch query to execute.
topicNoSearch category. The server applies `general` when this is absent. `news` automatically enables `include_published_date`.
countryNoBoost results from a country using Tavily's lowercase English country name (for example `united states`). Available only when `topic` is `general`.
endDateNoReturn results before this date (`YYYY-MM-DD`).
languageNoBoost or filter results by language — an ISO 639-1 code (for example `en`, `fr`, `zh-cn`) or English language name (for example `english`, `french`).
startDateNoReturn results after this date (`YYYY-MM-DD`).
timeRangeNoFilter by publish or last-updated date window.
exactMatchNoReturn only results containing the exact quoted phrase(s) in the query.
maxResultsNoMaximum search results to return. The server applies 10 when this is absent.
safeSearchNoFilter adult or unsafe content. Not supported when `search_depth` is `fast` or `ultra-fast`.
searchDepthNoLatency/relevance tradeoff. The server applies `basic` when this is absent. `advanced` costs 2 credits; `basic`, `fast` and `ultra-fast` cost 1 credit.
includeUsageNoInclude credit usage in the response.
includeAnswerNoInclude an LLM-generated answer. `true` or `basic` returns a quick answer; `advanced` returns a detailed answer. The server applies `false` when this is absent.
includeImagesNoInclude query-related images and per-result `images`.
autoParametersNoLet Tavily configure parameters from the query. Explicit values override auto-selected ones. `include_answer`, `include_raw_content` and `max_results` must always be set manually when using this.
excludeDomainsNoDomains to exclude (max 150).
includeDomainsNoDomains to include (max 300).
includeFaviconNoInclude a favicon URL per result.
chunksPerSourceNoMaximum relevant chunks per source in each result's `content`. The server applies 3 when this is absent. Available only when `search_depth` is `advanced`, `basic` or `fast`. Each chunk is at most 500 characters and joined with `[...]`.
filterByLanguageNoStrictly filter out non-matching languages. Requires `language`.
includeRawContentNoInclude cleaned page content per result. `true` or `markdown` returns markdown; `text` returns plain text and may increase latency. The server applies `false` when this is absent.
includeDomainsModeNoHow `include_domains` is applied. Requires `include_domains` to be set.
includePublishedDateNoInclude `published_date` on each result. Beta feature. Automatically enabled when `topic` is `news`.
filterByPublishedDateNoRemove results outside the date window or with no detectable date. Also enables `include_published_date`.
includeImageDescriptionsNoAdd descriptive text per image when `include_images` is true.

Schema Changelog

Changes observed during successful MCP inspections.

  1. First observedv0.1.0

TDQS

A3.8/5.0
Behavior3/5

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

Annotations declare readOnlyHint=false and openWorldHint=true, which is an unusual pairing for a read-only search and is left unexplained. The description adds a config recommendation (search_depth advanced, chunks_per_source 3) but doesn't clarify the credit costs or why a read-only operation isn't marked read-only.

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?

Three tight sentences: what it does, when to use it, and an actionable configuration tip. Front-loaded and zero filler.

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?

For a large but fully documented 25-parameter schema with no output schema, the description covers the essential purpose, usage trigger, and a key tuning recommendation. It's adequate, though the odd readOnlyHint=false on a search tool could have been addressed.

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%, so every one of the 25 parameters is already documented in the schema. The description names two parameters and their preferred values without adding semantics beyond that, so baseline 3 applies.

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

Purpose4/5

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

States a specific verb+resource ('Execute a real-time web search') and scopes it ('optimized for AI agents'). It distinguishes itself from tavily_research_create by being a real-time search vs. a research task, though it never explicitly names that sibling.

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?

Gives a clear usage trigger: 'Use when sources are unknown or current web context is needed.' This tells the agent when to reach for it over knowledge-only answers, though it doesn't name alternatives like firecrawl_scrape or tavily_research_create.

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

Deploy Server

Other Tools