Skip to content
TopicTracker
From HackerNewsView original
TranslationTranslation

Show HN: Knotic – layered memory (project/session/docs) for AI coding agents

Knotic introduces a layered memory system for AI coding agents, organizing context into project, session, and document layers to improve long-term coherence and reduce context loss during development.

Background

- **Knotic** is a new tool by Riccardo Tartaglia that gives AI coding assistants a structured memory organized in three layers: Project (persistent), Session (conversational), and Docs (reference). - **The problem it solves**: Current AI coding agents treat each conversation from scratch and quickly lose context. They forget project conventions, earlier decisions, or documentation they were told about even a few turns ago. - **Who it's for**: Developers using AI coding assistants who want the agent to remember project-wide context across sessions without constant repetition. - **Competing projects**: Mem0, Letta (formerly MemGPT), and Cursor's "rules" system also tackle memory. Knotic differentiates by offering explicit, user-defined layered memory rather than automatic recall.