Skip to main content
Glama
deepseek-mcp-server

deepseek-mcp-server

DeepSeek reasoner (R1 thinking mode)

deepseek_reason

Generates explicit chain-of-thought reasoning before answering hard math, logic, debugging, and multi-step planning prompts, returning the reasoning trace and final answer separately.

Instructions

Send a prompt to deepseek-reasoner, which produces an explicit chain of thought before its answer. Best for hard maths, logic, debugging and multi-step planning. Returns the reasoning trace and the final answer separately. Sampling parameters are ignored by this model and are dropped automatically.

Tool call template

This tool declares no required parameters. Add optional parameters from the input schema as needed.

{
  "id": 1,
  "jsonrpc": "2.0",
  "method": "tools/call",
  "params": {
    "arguments": {},
    "name": "deepseek_reason"
  }
}

Input Schema

NameRequiredDescriptionDefault
stopNoUp to 16 stop sequences. A single string is also accepted.
extraNoArbitrary extra request-body fields, merged last so they override everything else. The escape hatch for any API parameter this server does not model yet.
modelNoOverride the model id for this call.
toolsNoFunction-calling definitions, passed to the API verbatim. Each item is { type: 'function', function: { name, description, parameters } }.
top_pNoNucleus sampling, 0-1. Ignored by deepseek-reasoner.
promptNoThe user message. Use this for one-shot calls.
streamNoStream from the API and return the assembled result. Useful for very long outputs that would otherwise hit an idle timeout.
systemNoSystem instruction.
logprobsNoReturn log probabilities. deepseek-chat only.
messagesNoFull message array, OpenAI-compatible. Use this instead of `prompt` when you need multi-turn control. Roles: system, user, assistant, tool.
json_modeNoShorthand for response_format { type: 'json_object' }. Your prompt must mention JSON and should show an example.
max_tokensNoMaximum output tokens. Omit for the server-side maximum (8192 for deepseek-chat, 65536 for deepseek-reasoner). No client-side cap is applied.
session_idNoConversation key. Prior turns under this id are prepended automatically and this turn is appended, giving true multi-turn memory across tool calls.
timeout_msNoPer-call timeout override in milliseconds.
temperatureNo0.0-2.0. DeepSeek's own guidance: 0.0 for code and maths, 1.0 for general conversation and data analysis, 1.3 for creative writing. Ignored by deepseek-reasoner.
tool_choiceNo'auto', 'none', 'required', or { type: 'function', function: { name } }.
raw_responseNoReturn the complete unmodified API JSON instead of formatted text.
top_logprobsNo0-20, requires logprobs. deepseek-chat only.
include_usageNoAppend the token-usage and cache-hit footer. Default true.
presence_penaltyNo-2.0 to 2.0. Ignored by deepseek-reasoner.
frequency_penaltyNo-2.0 to 2.0. Ignored by deepseek-reasoner.
include_reasoningNoInclude the chain-of-thought block in the result. Default true for deepseek-reasoner.

Schema Changelog

Changes observed during successful MCP inspections.

  1. First observedv1.0.0

TDQS

A4/5.0
Behavior4/5

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

With zero annotations, the description carries the full load and does disclose the important quirks: the model emits a reasoning trace, returns trace and answer separately, and silently drops sampling parameters rather than erroring. It omits operational traits an agent would care about before a long call, such as added latency for reasoning or how reasoning tokens interact with max_tokens.

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?

Four short sentences with no filler: purpose first, then use-case fit, then return shape, then the sampling caveat. Nothing is repeated and every clause carries information an agent needs before calling.

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 22-parameter, annotation-free tool with no output schema, the description covers purpose, the unusual return structure, and the main behavioral trap, while the richly documented schema handles the remaining parameters. The gaps are minor: latency/cost of reasoning, and interaction between reasoning output and token limits.

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 coverage is 100%, so the baseline is 3, and the description adds only one cross-cutting parameter fact — that sampling parameters are ignored and dropped. That fact is already stated per-property in the schema for temperature, top_p, presence_penalty and frequency_penalty, so the description mostly restates structured data rather than adding new meaning.

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?

The description states a specific verb and target ("Send a prompt to deepseek-reasoner") and characterizes the model's distinctive output (an explicit chain of thought before the answer). The "best for hard maths, logic, debugging and multi-step planning" clause implicitly separates it from the general chat sibling, but no sibling is ever named or explicitly excluded, so differentiation stays inferential.

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?

It gives concrete when-to-use criteria (hard maths, logic, debugging, multi-step planning), which is enough for an agent to route work between this and a general chat tool. There are no explicit when-not-use conditions or named alternatives, so the agent must infer that simple or creative tasks belong elsewhere.

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