BigRack

About

BigRack is a Model Context Protocol (MCP) server designed to enhance AI development tools like Claude Code and Cursor. It addresses critical limitations faced by AI assistants in complex projects, such as progressive context loss, insufficient planning, and forgotten business rules. By providing a structured, persistent context and robust task management system, BigRack acts as a long-term memory for an AI assistant, enabling it to better manage business rules, architecture decisions, and task dependencies, ensuring more accurate and comprehensive development entirely locally.

Key Features

• Local-First Architecture: Ensures 100% local operation with data stored in SQLite and embedded vector search for privacy.
• MCP Protocol: Extends AI assistants with custom tools, compatible with Claude Code and Cursor.
• Task Decomposition: Breaks complex features into atomic tasks with automatic dependency graphs and intelligent recommendations.
• 2 GitHub stars
• Context Management: Stores business rules, glossary terms, and architecture patterns, accessible via semantic search and RAG.
• Vector Search & RAG: Utilizes local embeddings (Xenova/all-MiniLM-L6-v2) for natural language querying of project context.

Use Cases

• Solo Developers: Maintain context across long development sessions and plan complex features with confidence.
• Teams: Share business context consistently and track progress and dependencies in AI-assisted development.
• Enterprises: Manage complex projects with strict business rules, ensuring compliance and data privacy with self-hosted solutions.