DeepSeek Search Everywhere MCP

Overview Schema Source Related Servers Discussions Admin

deepseek_research

Gathers source material for a question with one content-rich search, then fetches missing or short page content from validated URLs and returns ranked sources with errors and truncation notes.

Instructions

Gather source material for a question: one search with page content, then fetch the gaps.

Runs a single search with need_content=True (no query rewrite, no model call, no summary). Results whose content is missing or shorter than 200 characters and whose URL passes validation are fetched directly with deepseek_fetch_many (at most 5 per call); a fetched page replaces the search content of that source. A URL that fails validation (for example an internal address) is not fetched. Failed or rejected sources keep the search content and set the source's error; the call still succeeds with partial=true.

Each source has id, title, url, authority, rerank_score, content, chars, origin ('search' or 'fetch'), error (null or {code, message}) and cached (whether its content came from the cache). Content is capped at per_page_chars per source and total_chars overall, in source order; a note says when any content was truncated. requests_used counts upstream requests made (the search plus each fetch; cache hits are free) and cached_count counts responses served from the cache. failed_count counts sources whose error is set and partial is true when it is non-zero. Synthesis is left to the caller.

Args: question: Search query, max 128 characters. Supports site:, -site:, "phrase", lang:. max_pages: Number of search results to use, 1-50 (default 5). per_page_chars: Content cap per source. total_chars: Total content budget across all sources. refresh: Bypass the response caches for the search and the fetches.

Input Schema

Name	Required	Description	Default
`refresh`	No
`question`	Yes
`max_pages`	No
`total_chars`	No
`per_page_chars`	No

Output Schema

Name	Required	Description	Default
No arguments

Schema Changelog

Changes observed during successful MCP inspections.

First observedv0.2.0Oct 27, 2026

TDQS

A4.5/5.0

Behavior5/5

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

With no annotations, the description carries the full behavioral burden and does so: it explains the fallback fetch path, the 200-character/URL-validation triggers, the 5-fetch cap, that failures keep search content and set error while the call still returns partial=true, cache-hit accounting in requests_used/cached_count, and truncation notes. Nothing important about failure modes or side effects is left implicit.

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

Conciseness3/5

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

It is front-loaded with the core action and organized into behavioral paragraphs followed by args, which is good. However, the sentence enumerating every source field (id, title, url, authority, rerank_score, content, chars, origin, error, cached) largely restates the output schema and adds bulk without new decision-relevant information.

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

Completeness5/5

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

For a 5-parameter tool with a multi-stage search-and-fetch pipeline and partial-failure semantics, the description covers trigger conditions, budget capping, caching, error propagation, and counting fields. Combined with an existing output schema, an agent has everything needed to call it correctly and interpret the result.

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

Parameters4/5

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

Schema description coverage is 0%, so the description must compensate, and it largely does: question gets a 128-character limit plus supported operators (site:, -site:, "phrase", lang:), max_pages gets the 1-50 range and default 5, and refresh is defined as bypassing caches for both search and fetches. per_page_chars and total_chars are only described as caps without ranges, which is the remaining gap.

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 opening sentence states a specific verb and resource plus the internal workflow: 'Gather source material for a question: one search with page content, then fetch the gaps.' It also disambiguates from siblings by declaring 'no query rewrite, no model call, no summary', which separates it from the search-and-read style tools listed alongside it.

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?

'Synthesis is left to the caller' establishes the boundary of responsibility, and the search-then-fetch-gaps framing tells an agent when this pipeline is the right shape. It stops short of explicitly naming deepseek_web_search or deepseek_search_and_read as alternatives with selection criteria, so no exclusion rules are given.

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

Deploy Server