011 GitHub stars
02Fully Local & Privacy-First processing for all research data
03Completely Customizable prompts, extraction schemas, and academic sources
04Automated Source Discovery with LLM-powered scraper for any URL and 7 pre-built plugins
05Hot-Loading Skill System with bundled and user-defined capabilities for dynamic agent behavior
06Letta-Powered Persistent Memory for agents to retain context and preferences across sessions