SDOF MCP - Structured Decision Optimization Framework

What is SDOF MCP - Structured Decision Optimization Framework?

SDOF MCP - Structured Decision Optimization Framework is a Model Context Protocol (MCP) server that provides persistent memory and context management for AI systems through a structured 5-phase optimization workflow. It is built for developers and AI practitioners who need to store, retrieve, and optimize knowledge with semantic search, vector embeddings, and multiple content types. It supports both MCP tools and an HTTP API, using OpenAI for embeddings and MongoDB or SQLite for storage.

How to use SDOF MCP - Structured Decision Optimization Framework?

Clone the repository, install dependencies with npm install, build with npm run build, set the OPENAI_API_KEY environment variable, and start the server with npm start. Then configure your MCP‑compatible client (e.g., Claude Desktop) to use the server with the stdio transport and the store_sdof_plan tool. Optionally, set environment variables like EMBEDDING_MODEL, HTTP_PORT, or MONGODB_URI.

Key features of SDOF MCP - Structured Decision Optimization Framework

• 5‑phase optimization workflow (Exploration, Analysis, Implementation, Evaluation, Integration)
• Vector embeddings with OpenAI for semantic search
• Persistent storage with MongoDB/SQLite and vector indexing
• Prompt caching optimized for LLM efficiency
• Schema validation for structured content types
• Multi‑interface: MCP tools and HTTP API (port 3000)

Use cases of SDOF MCP - Structured Decision Optimization Framework

• Storing decision records and rationales for architecture choices
• Saving code implementations and examples with metadata
• Capturing analysis results and findings for later retrieval
• Managing evaluation reports and performance metrics
• Consolidating integration documentation and learning artifacts

FAQ from SDOF MCP - Structured Decision Optimization Framework

What are the prerequisites for running SDOF MCP - Structured Decision Optimization Framework?

Node.js 18+ is required. You also need an OpenAI API key for embeddings and a MCP‑compatible client (such as Claude Desktop).

What databases are supported?

MongoDB and SQLite are supported for persistent storage with vector indexing. MongoDB URI is configured via the MONGODB_URI environment variable; by default SQLite is used.

What is the primary MCP tool provided?

The primary tool is store_sdof_plan. It accepts a plan_content string (Markdown) and a metadata object with fields like planTitle, planType, tags, phase, and cache_hint.

What content types can be stored?

Supported content types are: text, code, decision, analysis, solution, evaluation, and integration.

How do I configure the MCP client?

Add a stdio server entry to your MCP client configuration with the command node path/to/sdof-mcp/build/index.js and the OPENAI_API_KEY environment variable. Use alwaysAllow to auto‑approve the store_sdof_plan tool.