Injects raw chain-of-thought tokens from a reasoning model using Groq's API and Qwen's qwq model to improve performance in complex tasks.
Sponsored
This MCP server leverages Groq's API to expose raw chain-of-thought tokens from Qwen's qwq model. Inspired by Anthropic's research on the 'think' tool, it enables agents to use chain-of-thought reasoning as a scratchpad, enhancing performance on complex tool use situations. By instructing the AI to use this server on every request, the agent can list specific rules, verify information, and iterate over tool results for correctness before taking action.
主な機能
01Uses Groq's API to call LLMs.
02Improves AI agent performance in complex tasks.
03Exposes raw chain-of-thought tokens from reasoning models.
04Supports Qwen's qwq model.
05Integrates with MCP configurations.
ユースケース
01Enhancing AI agent's reasoning and problem-solving abilities.
02Facilitating chain-of-thought reasoning as a scratchpad for AI agents.
03Improving performance on tasks requiring complex tool use.