Tenets

Tenets is an open-source, privacy-first tool that solves the context problem for AI-assisted coding. It provides a Model Context Protocol (MCP) server, Python library, and CLI that automatically finds, ranks, and aggregates the most relevant code files for any development task. Instead of manually selecting files or feeding entire directories, Tenets uses multi-factor NLP ranking (BM25, TF-IDF, keyword extraction, import graphs, Git history) to deliver high-signal context to AI coding assistants while keeping your code entirely local. Targeted at developers and teams using AI coding assistants like Cursor, Claude Desktop, or Windsurf, Tenets integrates natively via MCP. You install it with a single pip command, configure your IDE, and Tenets exposes tools that the assistant calls on demand. It runs 100% on your machine—no cloud, no API keys needed for core features, making it ideal for privacy-conscious users and organizations. What sets Tenets apart from other MCP servers is its intelligent ranking layer. While others offer raw file access, Tenets analyzes code structure, import relationships, semantic meaning, and even Git signals to rank files by relevance. It includes optional ML embeddings (with tenets[ml]), token-budget-aware packing, and summarizers—both rules-based and ML—so the AI sees only the most important code within context limits. Beyond context building, Tenets offers project analysis tools: examining code complexity, hotspots, ownership, and team velocity. It also supports 'tenets'—guiding principles you can define and instill into sessions (e.g., 'Always validate user input') that persist across interactions. The CLI further enables dependency visualization, ranking queries, and session management.