Back to directory

llm-context-base

A git template for building a personal LLM-maintained wiki, implementing Andrej Karpathy's three-layer pattern of source documents, an LLM-maintained wiki and a schema config file.

llm-context-base

llm-context-base is a ready-to-use template for the LLM Wiki pattern described by Andrej Karpathy: raw source documents feed an LLM-maintained wiki, which follows a schema config file such as CLAUDE.md or AGENTS.md to stay consistent. Instead of designing that structure from scratch, a user clones the template, points an AI assistant at it, and starts talking; the wiki adapts to how the user works over time.

On top of the base pattern it adds a metadata standard, a queryable header of type, summary, tags, status and date on every file, so an LLM can scan headers instead of loading full files; an inbox-first capture area with a seven-day time-to-live; a training period that starts chatty and quiets as it adapts; a lint system that flags stale files, missing metadata and orphaned documents; and bootstrap shims for Claude Code, Cursor, Copilot, Windsurf and other tools.

Setup is cloning the template, opening it as an Obsidian vault, or any editor, alongside an AI tool, and filling in a starter config file.

More in Agent memory

See all 450
  • A long technical guide to AI agent memory: why stateless LLMs need it, the different memory architectures from vector stores to knowledge graphs, and how Cognee implements one as an open-source memory layer.

  • A local, file-based long-term memory layer for a personal AI that never summarizes conversations, keeping every line with a timestamp and searching by time first and words second, rather than retrieving only similar-seeming chunks.

  • NEOTH

    deepwiki.com

    A local-first personal AI daemon in Rust with five memory tiers, consent-gated tools, a WASM plugin sandbox, multi-provider model routing and signed audit logs, aimed at people who want both a daily assistant and a serious coding helper.