Think
Provides a lightweight 'think' tool for structured reasoning, enabling LLMs to pause, log thoughts, and improve multi-step problem
Think-MCP provides a simple 'think' tool for structured reasoning, inspired by Anthropic's Claude 'think' tool. This lightweight server implementation by Konstantin Krestnikov enables agentic LLMs to pause, log thoughts, and improve multi-step tool use without obtaining new information or changing databases. It's particularly useful for complex reasoning tasks where the model needs to work through a problem step-by-step before taking action.
Source
Repository: https://github.com/rai220/think-mcp
Maintain Think?
Let people know it's listed here — add the badge (live metrics, light/dark aware) or a plain link to your README or docs.
[Think on getagentictools](https://getagentictools.com/mcp/rai220-think-mcp?ref=badge)