wiki-to-graph

Turn an LLM wiki — a folder of interlinked markdown pages (the "LLM wiki" pattern by Andrej Karpathy) — into a real knowledge graph: typed nodes and edges you can validate, analyze, traverse, query, update, and view in a browser. Python 3 only; no other dependencies.

The insight: an LLM wiki is already a graph — pages are nodes, [[wiki-links]] are edges. Each page's consistent sections tell you what kind of edge each link is (a link under ## Related is a related edge; one under ## Contradictions is contradicts). This tool makes that graph explicit.

The included graph viewer (graph-viewer.html): nodes colored by kind, edges by type; click any node to read its summary and walk its edges.

This is a Claude plugin

wiki-to-graph/                      ← plugin root (also a one-plugin marketplace)
├── .claude-plugin/
│   ├── plugin.json                 ← plugin manifest
│   └── marketplace.json            ← lets the repo be added as a marketplace
├── skills/
│   └── wiki-to-graph/
│       ├── SKILL.md                ← the skill (build/validate/analyze/query/update/view)
│       ├── references/spec.md      ← full ontology + format spec
│       └── scripts/
│           ├── wiki_to_graph.py       ← the toolkit
│           └── build_graph_viewer.py   ← HTML graph viewer generator
├── examples/llm-wiki/              ← the runnable example wiki (source of build/)
│   ├── wiki/                       ← 28 markdown pages (the LLM wiki)
│   └── raw/                        ← 6 source papers the pages cite
├── build/                          ← sample outputs, regenerated from examples/llm-wiki/wiki
├── docs/outputs-and-workflows.md   ← what each build object is + example workflows
├── assets/graph-viewer.png
├── LICENSE.md
└── README.md

Runnable example included. examples/llm-wiki/ is the exact wiki the committed build/ artifacts were generated from, so the whole pipeline runs from a fresh clone. New here? Start with docs/outputs-and-workflows.md.

Install

• As a plugin (Cowork): open the delivered wiki-to-graph.plugin file and click install; or Settings → Capabilities → add plugin.
• As a marketplace / skill repo (Claude Code): /plugin marketplace add MangroveTechnologies/wiki-to-graph then /plugin install wiki-to-graph.
• From PyPI: pip install wiki-to-graph — installs the wiki-to-graph and wiki-to-graph-viewer console commands.
• No install needed: the scripts are plain Python — just run them (below).

Requirements: Python 3 (standard library only). networkx/scipy are optional, for your own heavier analysis. Distribution details and release steps: docs/publishing.md.

Quick start

Paths below are from the plugin root. (SCR=skills/wiki-to-graph/scripts)

1 · Build the graph

python3 skills/wiki-to-graph/scripts/wiki_to_graph.py build examples/llm-wiki/wiki \
        -o build/graph.json --emit sqlite,graphml

Writes build/graph.json (canonical), plus graph.db (SQLite) and graph.graphml (Gephi/yEd). Add --kst for a domain.json KST projection.

2 · Validate

python3 skills/wiki-to-graph/scripts/wiki_to_graph.py validate build/graph.json

Broken links / orphans / self-loops fail (exit 1). Cross-reference cycles are informational.

3 · Analyze

python3 skills/wiki-to-graph/scripts/wiki_to_graph.py analyze build/graph.json --top 5

PageRank, in-degree, most-contested nodes, components, communities. Shortest path: --path "GPT-3" "Layer Normalization".

4 · Query (no SQL needed)

SCR=skills/wiki-to-graph/scripts/wiki_to_graph.py
python3 $SCR query build/graph.json node "RLHF"            # details + edges
python3 $SCR query build/graph.json neighbors "GPT-3"      # outgoing
python3 $SCR query build/graph.json backlinks "Transformer"# incoming
python3 $SCR query build/graph.json contradicts            # all tension pairs
python3 $SCR query build/graph.json bfs "Transformer" --edges related
python3 $SCR query build/graph.json dfs "GPT-3" --edges contradicts --undirected
python3 $SCR query build/graph.json path "Positional Encoding" "RLHF"

Filter any traversal on edge type, node type/kind, or a combination — include or exclude:

5 · Update the wiki, then rebuild

The graph is derived; edit the source markdown and re-run build.

python3 $SCR update examples/llm-wiki/wiki add-node --title "Mixture of Experts" --kind schema --summary "…"
python3 $SCR update examples/llm-wiki/wiki add-edge --from "Mixture of Experts" --to "Transformer" --type related
python3 $SCR update examples/llm-wiki/wiki set-kind --node "GPT-3" --kind schema

6 · View in a browser

python3 skills/wiki-to-graph/scripts/build_graph_viewer.py build/graph.json -o build/graph-viewer.html

Double-click build/graph-viewer.html (offline, no dependencies).

The model in 30 seconds

• Nodes have a structural type (concept, source, index, log); concepts also carry a knowledge kind: concept / schema / procedure / fact (set per page via frontmatter kind:).
• Edges are typed by their source section: mentions, related, contradicts, cites, plus indexes / records from the index/log hub pages.
• Each node carries its own edges list, degrees, word_count, n_sources, aliases. Link text is stored as plain names — the relationship lives in the edge, not in [[markup]].

Full details: skills/wiki-to-graph/references/spec.md.

Use on your own wiki

One concept per page, consistent ## sections, [[Page Title]] links, optional kind: frontmatter. Different section names? Pass --map map.json to build.

License

Licensed under Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0). Free to use, share, and adapt for non-commercial purposes, with attribution, under the same license. Commercial use is not permitted under this license.

Commercial users should inquire for use: contact tim.darrah@mangrove.ai.

Full terms in LICENSE.md.

Status & limits

Working end-to-end: build → validate → analyze → query → update → view. Deliberately simple and static — the graph is recomputed from the markdown on every build (no incremental updates). Edge weight is captured but inert (not used by metrics). Not yet aligned to any external ontology.

Acknowledgements

• The "LLM wiki" pattern is due to Andrej Karpathy.
• The bundled example (examples/llm-wiki/) follows Data Science Dojo's tutorial, The LLM Wiki Pattern by Andrej Karpathy: A Step-by-Step Tutorial to Building a Compounding Knowledge Base, and is compiled from six foundational AI papers (Attention Is All You Need, BERT, GPT-3, Foundation Models, RLHF/InstructGPT, Chinchilla).