LLM Wiki Tools — No-Code & Open-Source Directories
This is the LLM Wiki tools hub — the entry point to two curated directories of every serious implementation of Karpathy's LLM Wiki pattern we have actually installed and reviewed. Pick the path that matches how you work. New to the pattern itself? The plain-English primer takes 5 minutes.
Two directories, two audiences
We split the LLM Wiki tools landscape into two directories so you do not have to wade through tools that are not for you:
🧰 No-Code LLM Wiki Tools
For: writers, consultants, researchers, and anyone whose day job keeps them out of a terminal.
Six LLM Wiki tools you can run with zero code, ranked by how fast a non-developer can get a useful page out of them. Includes desktop apps, web apps, Obsidian plugins, and Notion-native paths. Every tool has been installed and tested before review.
💻 Open-Source LLM Wiki Implementations
For: developers and AI engineers who want to read the source, fork the pattern, or build something custom.
Fifteen GitHub repos (and counting), each one installed and verified before listing. Includes the original Karpathy gist, Claude Code plugins, Obsidian + Ollama bundles, persistent-memory engines, and more. Filter by stack, difficulty, and best-fit use case.
Browse open-source implementations →
Which one should I read first?
| If you… | Start with |
|---|---|
| Already use Obsidian, Notion, or refuse to install a terminal app | No-Code Tools |
| Live inside Claude Code, Cursor, or a CLI editor | Open-Source Implementations |
| Are not sure yet | Read the primer first, then come back here |
Why a "tools directory" at all?
When Karpathy's LLM Wiki gist went viral in early April 2026, GitHub filled up with a dozen serious implementations and three dozen experiments within a single week. None of the existing "AI tools" directories cover this niche, and the README files of most repos optimistically describe features that are not yet shipped. We exist to fill that gap with honest, hands-on reviews.
Every tool on either directory passes three checks before it gets listed:
- Installable — we cloned, ran, and got it past hello-world.
- Schema-aware — the implementation actually uses the LLM Wiki concept of a schema, not just "summarize the docs into wiki pages."
- Honest verdict — at least 30 minutes of real use before writing the verdict.
Want a tested setup that just works?
If browsing 21 tools sounds like more work than it should be, our LLM Wiki Templates page ships drop-in vaults for Obsidian and Notion plus a full Starter Kit with CLAUDE.md, schemas, and the ingest pipeline pre-built. One purchase, lifetime updates, no schema-tuning weekend required.
Related reading
- What is LLM Wiki? — the conceptual primer
- LLM Wiki vs RAG — when to use which architecture
- Karpathy's LLM Wiki with Claude Code: Setup Tutorial — build your own from scratch
- Free LLM Wiki schemas — three production schemas to drop into any tool above