Automates long-term memory and knowledge extraction to improve Claude's performance across coding sessions.
Nous provides a sophisticated self-improvement pipeline for Claude Code by capturing behavioral insights and project-specific facts through automated session analysis. By monitoring context usage via a lightweight statusline integration, it identifies optimal moments to extract 'Learnings' (workflow patterns) and 'Knowledge' (architectural details), which are then injected into the context of future sessions. This creates a persistent memory system that helps Claude avoid repeating mistakes and maintain deep project awareness without requiring manual documentation or context management from the developer.
Key Features
01Dual-lens processing for behavioral learnings and factual project knowledge
02Lightweight statusline integration for activity logging and performance monitoring
030 GitHub stars
04Automated session extraction based on context window thresholds
05Seamless context injection at the start of every new development session
06Persistent per-project storage using structured engram and cortex data files
Use Cases
01Reducing repetitive prompting by persisting architectural decisions and file locations
02Maintaining development continuity across long-term, complex projects
03Capturing and enforcing custom coding standards or preferred tool patterns automatically
