"The concept of thinking or introspection" matching MCP servers:

Simple sequential thinking MCP in python

A Model Context Protocol (MCP) server implementation for the Google Gemini language model. This server allows Claude Desktop users to access the powerful reasoning capabilities of Gemini-2.0-flash-thinking-exp-01-21 model.

Enables AI assistants to work through complex problems step-by-step with dynamic thought processes, allowing for revision of previous steps, exploration of alternative approaches, and flexible planning as understanding deepens.

Enables structured, step-by-step problem-solving through dynamic thinking processes that can be revised, branched, and adjusted as understanding deepens. Supports breaking down complex problems into manageable steps with the ability to revise previous thoughts and explore alternative reasoning paths.

An MCP server that extends AI coding assistants with deterministic, algorithmic capabilities such as code analysis, fault localization, and formal verification, enabling an autonomous engineering team within the IDE.

An enhanced sequential thinking tool optimized for programming tasks that helps break down complex coding problems into structured, self-auditing thought steps with branching and revision capabilities.

Enables parallel execution of shell commands and AI coding agents (Claude, Gemini, Codex) across lists of items like files or URLs, with batched processing and real-time streaming output for batch operations.

A robust, language-agnostic Model Context Protocol (MCP) server that provides AI coding agents with the ability to edit files surgically via Abstract Syntax Trees (AST) instead of relying on token-heavy, brittle search-and-replace or diff operations.

MCP server for DeepSeek AI models (Chat + Reasoner). Supports multi-turn sessions, model fallback with circuit breaker, function calling, thinking mode, JSON output, multimodal input, and cost tracking.

Provides per-Subagent MCP controls to any coding agent or client across all your MCPs and prevents context window waste. Loads only 3 tools instead of all your MCP Server's tool definitions. Agents discover tools on-demand, only when needed and only the servers and tools they are allowed.

Connects AI coding agents to the SkillsMP marketplace, allowing users to search, read, and install over 8,000 community-made skills. It enables agents to gain new capabilities either through on-the-spot instruction or permanent installation without requiring an API key.

Connects AI coding assistants to Roblox Studio, enabling direct manipulation of the DataModel through natural language commands with undo support.

Enables a supervisor LLM to delegate coding tasks to a persistent OpenCode server, block until the turn completes or needs input, and resume the same turn with answers or permission decisions.

Delegates work from MCP clients (like Claude Code) to the Codex CLI, allowing spawning of autonomous Codex subagents for tasks.

Runs as an MCP server providing file system operations, shell execution, and system information tools, allowing Claude or any MCP client to edit files and control the machine.

Enables Claude Code to run OpenAI Codex (GPT-5) tasks asynchronously in the background, filtering out thinking logs to save 95% context tokens and allowing parallel execution.

Self-hosted coding agent that runs Claude Code or Codex in your own infrastructure and works from Slack. Exposes codebase investigation, file edits, branch and PR creation, allowlisted command execution, CI status and an audit log as MCP tools, keeping every credential behind the tool layer instead of handing it to the model.

MCP bridge server for patchwork-os. Boots a stdio-HTTP bridge with MCP introspection over /mcp, supporting tools/list and the full MCP protocol over HTTP.

Compare LLM inference costs and find cheap alternatives to OpenAI/Anthropic/DeepSeek. Gonka Network offers an OpenAI-compatible API at up to 6800x lower cost than GPT-4o.

Wraps the cursor-agent CLI to provide cost-effective tools for repository analysis, code search, planning, and editing. Offloads heavy thinking tasks from the host AI to reduce token usage while maintaining precise, scoped workspace operations.