Overview
What is Tianji Thinking Models?
Tianji Thinking Models is an MCP server that integrates hundreds of thinking models, frameworks, and methodologies to help users think more systematically about problems. It enables AI assistants to access these tools through the Model Context Protocol (MCP) and apply structured thinking methods during conversations.
How to use Tianji Thinking Models?
The server requires Node.js 18+ and TypeScript 4.9+. After installation, configure it as an MCP server to connect AI assistants, then use its 20+ tools for exploring, recommending, interactively reasoning, and creating thinking models. Supported languages are Chinese (zh) and English (en).
Key features of Tianji Thinking Models
- Rich library of hundreds of thinking models across multiple domains
- Intelligent model recommendations based on problem characteristics
- Interactive reasoning process guidance with step-by-step analysis
- Learning and adaptation system that improves recommendations via feedback
- Model creation and combination for innovative thinking frameworks
- Comprehensive tool set for exploration, problem-solving, creation, and system learning
Use cases of Tianji Thinking Models
- Explore and search a vast library of systematic thinking models for any problem
- Receive tailored model recommendations and interactive reasoning guidance for complex challenges
- Generate multiple hypotheses and validate them using structured methods
- Create new thinking models or combine existing ones for custom frameworks
- Analyze knowledge gaps and track usage statistics to improve thinking skills
FAQ from Tianji Thinking Models
What runtime does Tianji Thinking Models require?
It requires Node.js 18+ and TypeScript 4.9+, with MCP Protocol 1.11+ and Zod 3.24+.
What languages are supported?
The server supports Chinese (zh) and English (en) for all tool interactions and model data.
Can I create my own thinking models?
Yes, the server provides create-thinking-model, update-thinking-model, emergent-model-design, and delete-thinking-model tools for custom model creation and management.
Does the server learn from user feedback?
Yes, the record-user-feedback tool allows users to provide feedback on model experiences, and the recommendation system can apply learning adjustments (controlled by the use_learning_adjustment parameter).
How many thinking models are included?
The server contains hundreds of thinking models across multiple categories and subcategories; the total count can be retrieved via the count-models tool.
