Skip to content
Back to search

mindbase-llm-wiki

Karpathy's LLM Wiki idea as a product — an AI that builds and maintains a markdown wiki from your notes and sources. MCP server + web UI, runs on free local models (Ollama), no API key needed. MIT.

Safety scan

A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.

Checked today
Badge for your README
findsafeskills safety rating

Markdown

[![findsafeskills safety rating](/api/listings/frankchu91--mindbase-llm-wiki/badge.svg)](/listings/frankchu91--mindbase-llm-wiki)

HTML

<a href="/listings/frankchu91--mindbase-llm-wiki"><img src="/api/listings/frankchu91--mindbase-llm-wiki/badge.svg" alt="findsafeskills safety rating" /></a>

Install

i

Install instructions from the README

As written by the repo's author. findsafeskills has not verified these steps.

**Browser (fastest):** `npx mindbase-app` — shown above. Everything runs locally at `localhost:4321`.

**Claude Code (flagship):** the full Karpathy 8-step ingest with sub-agents, slash commands, and per-agent tool boundaries:

/plugin marketplace add frankchu91/mindbase-llm-wiki
/plugin install mb@mindbase

Restart when prompted, then type `/` — you should see `/mb:contribute`, `/mb:build`, `/mb:ask`, `/mb:lint` and 8 more. You get 5 sub-agents with strict tool allowlists (the builder has no file-write tool at all — only an atomic-write MCP call), plus a SessionStart hook that auto-injects your project context.

<p align="center"> </p>

<details> <summary><b>Cursor</b></summary>

Add to `~/.cursor/mcp.json`:

{ "mcpServers": { "mindbase": { "command": "npx", "args": ["-y", "mindbase-mcp"] } } }

Restart Cursor — the tool picker should list `mindbase_contribute` and 48 others. Recommended: add a conventions block to `~/.cursor/rules.md` so the LLM reliably routes "add to mindbase X" to the tools — copy it from the [guide](docs/guide.md#the-llm-in-cursorwindsurf-ignores-my-add-to-mindbase-request).

</details>

<details> <summary><b>Windsurf</b></summary>

Cascade settings → MCP → add:

{ "mcpServers": { "mindbase": { "command": "npx", "args": ["-y", "mindbase-mcp"] } } }

Same rules-file approach as Cursor works for Cascade.

</details>

<details> <summary><b>Cline (VSCode)</b></summary>

Cline settings → **MCP Servers**:

{ "mcpServers": { "mindbase": { "command": "npx", "args": ["-y", "mindbase-mcp"] } } }

Cline auto-detects; every tool call gets a confirmation dialog by default.

</details>

<details> <summary><b>Continue.dev</b></summary>

`~/.continue/config.json`, under `experimental.modelContextProtocolServers`:

{ "experimental": { "modelContextProtocolServers": [ { "transport": { "type": "stdio", "command": "npx", "args": ["-y", "mindbase-mcp"] } } ] } }

MCP tools appear under `@` in chat.

</details>

<details> <summary><b>Any other MCP client</b> (Zed, Aider, Goose, Claude Agent SDK…)</summary>

Point it at `npx -y mindbase-mcp` as a stdio server. See your client's MCP docs for the config location.

</details>

**Next:** create your first project and learn the four daily workflows in the **[Guide →](docs/guide.md)**

What was scanned

The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.

.claude-plugin/marketplace.json · Manifest · 23 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

  1. {
  2. "$schema": "https://anthropic.com/claude-code/marketplace.schema.json",
  3. "name": "mindbase",
  4. "description": "MindBase: Karpathy's LLM-Wiki pattern, no terminal required.",
  5. "owner": {
  6. "name": "Haobing Chu",
  7. "email": "haobing0304@gmail.com"
  8. },
  9. "plugins": [
  10. {
  11. "name": "mb",
  12. "description": "Build and maintain a persistent AI-tended research wiki. Includes MCP server with 49 tools, slash commands, sub-agents, and an auto-context-injection hook.",
  13. "version": "0.4.3",
  14. "author": {
  15. "name": "Haobing Chu",
  16. "email": "haobing0304@gmail.com"
  17. },
  18. "source": "./apps/plugin",
  19. "category": "productivity",
  20. "homepage": "https://github.com/frankchu91/mindbase#readme"
  21. }
  22. ]
  23. }
README.md · README · 232 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

  1. # MindBase — Karpathy's LLM Wiki, as a product
  2. [![npm](https://img.shields.io/npm/v/mindbase-mcp?label=mindbase-mcp)](https://www.npmjs.com/package/mindbase-mcp)
  3. [![npm downloads](https://img.shields.io/npm/dw/mindbase-mcp)](https://www.npmjs.com/package/mindbase-mcp)
  4. [![CI](https://github.com/frankchu91/mindbase-llm-wiki/actions/workflows/ci.yml/badge.svg)](https://github.com/frankchu91/mindbase-llm-wiki/actions)
  5. [![License: MIT](https://img.shields.io/badge/license-MIT-green)](LICENSE)
  6. [![Website](https://img.shields.io/badge/website-live%20demos-blue)](https://frankchu91.github.io/mindbase-llm-wiki/)
  7. [![Glama score](https://glama.ai/mcp/servers/frankchu91/mindbase/badges/score.svg)](https://glama.ai/mcp/servers/frankchu91/mindbase)
  8. > **An open-source implementation of Andrej Karpathy's LLM Wiki idea: an AI that builds and maintains a wiki from your sources.** Not RAG-in-a-vector-DB. A real markdown wiki on your disk, that an LLM gardens for you between conversations.
  9. ```bash
  10. npx mindbase-app
  11. ```
  12. One command: starts the local server, opens the web app, and walks you through picking a **free local model** that fits your RAM. No API key, nothing leaves your machine. (Node 20+)
  13. <p align="center">
  14. <img alt="Demo: /contribute in the MindBase web UI — the AI shows takeaways and a checkbox plan, and writes to the LLM wiki only after approval" src="docs/assets/contribute.gif" width="880">
  15. </p>
  16. MindBase implements Andrej Karpathy's [LLM-Wiki pattern](https://x.com/karpathy/status/1911080091498963196): you feed it sources (papers, articles, thoughts); the LLM reads, cross-references, flags contradictions, and writes structured wiki pages. Later, when you ask a question, the wiki already has the synthesized answer — no vector-search re-derivation at query time.
  17. **Status:** Early access, actively developed. What's new: [CHANGELOG](CHANGELOG.md) · [Releases](https://github.com/frankchu91/mindbase-llm-wiki/releases)
  18. ---
  19. ## Why MindBase
  20. You read a lot. Papers, articles, tweets, docs. You want to remember them, connect them, form opinions from them. Today you have two bad options:
  21. - **Notion / Obsidian / Roam:** Passive containers. You do all the organizing. AI features are bolted-on generation, not maintenance.
  22. - **NotebookLM / Perplexity Pages / ChatGPT search:** RAG-based. Nothing accumulates. Every question re-derives the answer from raw sources.
  23. **MindBase is the third option:** the LLM actively maintains a persistent, structured wiki as you feed it sources. Knowledge compounds. Your `context.md` gets sharper every time you contribute. The AI *remembers you across sessions* because your beliefs are written down in markdown files — not stored in a chat history that gets summarized away.
  24. Think of it as **a personal Wikipedia that an AI intern writes for you**, kept up to date, cross-referenced, and honest about what it doesn't know.
  25. ## How it works (30 seconds)
  26. Three physical layers on disk (Karpathy's model):
  27. <p align="center">
  28. <picture>
  29. <source media="(prefers-color-scheme: dark)" srcset="docs/assets/hero-dark.svg">
  30. <img alt="MindBase: you feed sources, the LLM gardens them into a wiki that compounds — plain markdown on your disk" src="docs/assets/hero-light.svg" width="920">
  31. </picture>
  32. </p>
  33. | Layer | Who owns it | What lives there |
  34. |---|---|---|
  35. | `sources/` | **You** — append-only, the AI never rewrites it | Quick captures, full notes, PDFs, URLs |
  36. | `context.md` + `sources/research/` | **The AI** — every change human-approved | The maintained wiki: synthesis, concept pages, `[[wikilinks]]` |
  37. | `state/` · `logs/` · `artifacts/` | Derived — always rebuildable | Search index, snapshots, lint findings, op history |
  38. Three operations run the loop: **ingest** (AI reads a source, discusses takeaways, you approve the wiki updates), **build** (regenerate `context.md` from everything unbuilt), **lint** (the AI audits its own wiki for contradictions, stale claims, and orphans).
  39. ## What makes it different
  40. - **The AI asks before writing.** Every ingest shows takeaways + a checkbox plan; only what you approve gets written. No black-box edits to your knowledge.
  41. - **You can watch the wiki absorb your notes.** Every note carries a status chip — ✨ *Add to wiki* until digested, ✓ *In wiki* after.
  42. - **It audits its own knowledge.** One command re-reads the whole wiki and reports contradictions with the exact conflicting sentences quoted. Notion and NotebookLM structurally cannot do this.
  43. - **Free and local by default.** Hardware-detect wizard installs the best Ollama model for your RAM. Cloud keys optional.
  44. - **Plain markdown on disk.** Grep it, git it, open it in Obsidian, leave anytime.
  45. ## Install
  46. **Browser (fastest):** `npx mindbase-app` — shown above. Everything runs locally at `localhost:4321`.
  47. **Claude Code (flagship):** the full Karpathy 8-step ingest with sub-agents, slash commands, and per-agent tool boundaries:
  48. ```
  49. /plugin marketplace add frankchu91/mindbase-llm-wiki
  50. /plugin install mb@mindbase
  51. ```
  52. Restart when prompted, then type `/` — you should see `/mb:contribute`, `/mb:build`, `/mb:ask`, `/mb:lint` and 8 more. You get 5 sub-agents with strict tool allowlists (the builder has no file-write tool at all — only an atomic-write MCP call), plus a SessionStart hook that auto-injects your project context.
  53. <p align="center">
  54. <img alt="The core flow in Claude Code: /mb:contribute with an arXiv link — download, archive, discuss takeaways, approve, commit" src="docs/assets/terminal-contribute.png" width="880">
  55. </p>
  56. <details>
  57. <summary><b>Cursor</b></summary>
  58. Add to `~/.cursor/mcp.json`:
  59. ```json
  60. { "mcpServers": { "mindbase": { "command": "npx", "args": ["-y", "mindbase-mcp"] } } }
  61. ```
  62. Restart Cursor — the tool picker should list `mindbase_contribute` and 48 others. Recommended: add a conventions block to `~/.cursor/rules.md` so the LLM reliably routes "add to mindbase X" to the tools — copy it from the [guide](docs/guide.md#the-llm-in-cursorwindsurf-ignores-my-add-to-mindbase-request).
  63. </details>
  64. <details>
  65. <summary><b>Windsurf</b></summary>
  66. Cascade settings → MCP → add:
  67. ```json
  68. { "mcpServers": { "mindbase": { "command": "npx", "args": ["-y", "mindbase-mcp"] } } }
  69. ```
  70. Same rules-file approach as Cursor works for Cascade.
  71. </details>
  72. <details>
  73. <summary><b>Cline (VSCode)</b></summary>
  74. Cline settings → **MCP Servers**:
  75. ```json
  76. { "mcpServers": { "mindbase": { "command": "npx", "args": ["-y", "mindbase-mcp"] } } }
  77. ```
  78. Cline auto-detects; every tool call gets a confirmation dialog by default.
  79. </details>
  80. <details>
  81. <summary><b>Continue.dev</b></summary>
  82. `~/.continue/config.json`, under `experimental.modelContextProtocolServers`:
  83. ```json
  84. { "experimental": { "modelContextProtocolServers": [ { "transport": { "type": "stdio", "command": "npx", "args": ["-y", "mindbase-mcp"] } } ] } }
  85. ```
  86. MCP tools appear under `@` in chat.
  87. </details>
  88. <details>
  89. <summary><b>Any other MCP client</b> (Zed, Aider, Goose, Claude Agent SDK…)</summary>
  90. Point it at `npx -y mindbase-mcp` as a stdio server. See your client's MCP docs for the config location.
  91. </details>
  92. **Next:** create your first project and learn the four daily workflows in the **[Guide →](docs/guide.md)**
  93. ## The web UI
  94. Since 0.3 the browser app stands on its own — write notes in a full WYSIWYG editor (`Cmd+N`), quick-capture from anywhere (`Cmd+I`), and run the AI operations with approval cards: `/contribute`, `/build`, `/lint`, `/research`. Live demos on the **[website](https://frankchu91.github.io/mindbase-llm-wiki/)**.
  95. <p align="center">
  96. <img alt="MindBase web UI — category tree on the left, the LLM-maintained context.md in the center, chat starters on the right" src="docs/assets/webui.png" width="920">
  97. </p>
  98. **Free local models:** the setup wizard detects your hardware and installs what fits — `llama3.2:3b` (8GB), `qwen3:14b` (24GB+), or Meta's **Muse Glimmer 30B** (32GB+ Apple Silicon, Ollama ≥ 0.32.7). Measured guidance: qwen3:14b for interactive work (~30s), Glimmer for background lint/build — slower, but its findings quote the exact conflicting sentences. The model switcher on the chat composer flips between them in two clicks.
  99. <details>
  100. <summary><b>Feature matrix by editor</b></summary>
  101. | Feature | Claude Code | Cursor / Windsurf / Cline / Continue | Web UI |
  102. |---|---|---|---|
  103. | Slash commands (`/mb:*`) | ✅ | ❌ (use natural language) | ✅ (`/contribute`, `/build`, `/lint`, `/research`) |
  104. | MCP tools directly | ✅ | ✅ | ❌ |
  105. | Karpathy 8-step ingest with approval | ✅ (sub-agents) | ⚠️ Manual via prompt | ✅ (approval cards) |
  106. | Contribute / ingest PDF & URL | ✅ | ✅ | ✅ (upload + ✨ Process) |
  107. | Ask wiki with cited answers | ✅ | ✅ | ✅ |
  108. | Build / health check | ✅ | ✅ | ✅ |
  109. | Wiki tree browsing + rich editor | ❌ | ❌ | ✅ |
  110. | `-p project-id` routing | ✅ | ⚠️ natural language | ✅ (switcher) |
  111. </details>
  112. ## Where your data lives
  113. Everything is plain markdown under `~/mindbase-data/` (override: `MINDBASE_DATA_DIR`):
  114. ```
  115. ~/mindbase-data/projects/my-research/
  116. ├── README.md # Ops manual — you edit, LLM reads
  117. ├── context.md # Synthesized truth — LLM writes, you approve
  118. ├── index.yaml # Auto-generated catalog
  119. ├── sources/
  120. │ ├── contributors/<you>/ # Your dated entries + notes (append-only)
  121. │ ├── research/ # LLM-authored wiki pages
  122. │ └── raw/ # PDFs, HTML captures
  123. ├── logs/ # Chronological operation log
  124. ├── artifacts/ # Briefs, exports, lint findings
  125. └── state/builder/snapshots/ # context.md snapshots for rollback
  126. ```
  127. No proprietary database — what you see on disk is what MindBase knows. `git init` it, back it up with anything, delete a project with `rm -rf`.
  128. ## Architecture at a glance
  129. ```
  130. Claude Code / Cursor / any MCP editor Web UI (npx mindbase-app)
  131. │ MCP · sub-agents with │ /commands ·
  132. │ per-agent tool allowlists │ approval cards
  133. └─────────────┬─────────────────────────┬──────┘
  134. ▼ ▼
  135. ┌──────────────────────────────────────────────┐
  136. │ One ops engine: gather context → single │
  137. │ constrained JSON completion → human │
  138. │ approval → whitelisted executors → log │
  139. └─────────────────────┬────────────────────────┘
  140. ▼
  141. ~/mindbase-data/ · plain markdown
  142. (LLM: Ollama local models or any cloud key)
  143. ```
  144. Monorepo: `packages/core` (TS strict library) · `apps/mcp` (49-tool MCP server) · `apps/server` + `apps/web` (Express + React UI) · `apps/app` (npx launcher) · `apps/plugin` (Claude Code bundle).
  145. ## Docs & help
  146. - **[Guide](docs/guide.md)** — first project, the four daily workflows, multi-project routing, troubleshooting
  147. - **[Website](https://frankchu91.github.io/mindbase-llm-wiki/)** — live demos
  148. - **[CHANGELOG](CHANGELOG.md)** · **[Issues](https://github.com/frankchu91/mindbase-llm-wiki/issues)** — I reply to every issue same-day during beta
  149. - **The idea:** [Karpathy's LLM Wiki gist](https://gist.github.com/karpathy/442a6bf555914893e9891c11519de94f)
  150. ## Roadmap
  151. **Next:** browser extension for one-click page capture · audio input via Whisper · unified meta-tool for Cursor/Windsurf slash-like UX.
  152. **Later:** team projects with human-in-the-loop review · audio digests · desktop app · mobile capture.
  153. ## Feedback
  154. Beta through 2026-Q4. If you tried MindBase and gave up — please tell me why: [issues](https://github.com/frankchu91/mindbase-llm-wiki/issues) or [haobing0304@gmail.com](mailto:haobing0304@gmail.com). The blockers you hit are gold.
  155. ## License
  156. [MIT](LICENSE) — do what you want, no warranty. If you build something interesting on top, I'd love to hear about it.
  157. ---
  158. **Built with the belief that AI's most valuable gift is not "generation on demand" but "gardening of a persistent artifact you own."**

Ratings and comments

Sign in to rate this skill.

Sign in

Free to use. If it’s helped, pass it on.