{
"markdown": "# MindBase — Karpathy's LLM Wiki, as a product\n\n[](https://www.npmjs.com/package/mindbase-mcp)\n[](https://www.npmjs.com/package/mindbase-mcp)\n[](https://github.com/frankchu91/mindbase-llm-wiki/actions)\n[](LICENSE)\n[](https://frankchu91.github.io/mindbase-llm-wiki/)\n[](https://glama.ai/mcp/servers/frankchu91/mindbase)\n\n> **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.\n\n```bash\nnpx mindbase-app\n```\n\nOne 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+)\n\n
\n \n
\n\nMindBase 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.\n\n**Status:** Early access, actively developed. What's new: [CHANGELOG](CHANGELOG.md) · [Releases](https://github.com/frankchu91/mindbase-llm-wiki/releases)\n\n---\n\n## Why MindBase\n\nYou read a lot. Papers, articles, tweets, docs. You want to remember them, connect them, form opinions from them. Today you have two bad options:\n\n- **Notion / Obsidian / Roam:** Passive containers. You do all the organizing. AI features are bolted-on generation, not maintenance.\n- **NotebookLM / Perplexity Pages / ChatGPT search:** RAG-based. Nothing accumulates. Every question re-derives the answer from raw sources.\n\n**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.\n\nThink 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.\n\n## How it works (30 seconds)\n\nThree physical layers on disk (Karpathy's model):\n\n
\n \n \n \n \n
\n\n| Layer | Who owns it | What lives there |\n|---|---|---|\n| `sources/` | **You** — append-only, the AI never rewrites it | Quick captures, full notes, PDFs, URLs |\n| `context.md` + `sources/research/` | **The AI** — every change human-approved | The maintained wiki: synthesis, concept pages, `[[wikilinks]]` |\n| `state/` · `logs/` · `artifacts/` | Derived — always rebuildable | Search index, snapshots, lint findings, op history |\n\nThree 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).\n\n## What makes it different\n\n- **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.\n- **You can watch the wiki absorb your notes.** Every note carries a status chip — ✨ *Add to wiki* until digested, ✓ *In wiki* after.\n- **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.\n- **Free and local by default.** Hardware-detect wizard installs the best Ollama model for your RAM. Cloud keys optional.\n- **Plain markdown on disk.** Grep it, git it, open it in Obsidian, leave anytime.\n\n## Install\n\n**Browser (fastest):** `npx mindbase-app` — shown above. Everything runs locally at `localhost:4321`.\n\n**Claude Code (flagship):** the full Karpathy 8-step ingest with sub-agents, slash commands, and per-agent tool boundaries:\n\n```\n/plugin marketplace add frankchu91/mindbase-llm-wiki\n/plugin install mb@mindbase\n```\n\nRestart 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.\n\n
\n \n
\n\n\nCursor\n\nAdd to `~/.cursor/mcp.json`:\n\n```json\n{ \"mcpServers\": { \"mindbase\": { \"command\": \"npx\", \"args\": [\"-y\", \"mindbase-mcp\"] } } }\n```\n\nRestart 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).\n\n\n\n\nWindsurf\n\nCascade settings → MCP → add:\n\n```json\n{ \"mcpServers\": { \"mindbase\": { \"command\": \"npx\", \"args\": [\"-y\", \"mindbase-mcp\"] } } }\n```\n\nSame rules-file approach as Cursor works for Cascade.\n\n\n\n\nCline (VSCode)\n\nCline settings → **MCP Servers**:\n\n```json\n{ \"mcpServers\": { \"mindbase\": { \"command\": \"npx\", \"args\": [\"-y\", \"mindbase-mcp\"] } } }\n```\n\nCline auto-detects; every tool call gets a confirmation dialog by default.\n\n\n\n\nContinue.dev\n\n`~/.continue/config.json`, under `experimental.modelContextProtocolServers`:\n\n```json\n{ \"experimental\": { \"modelContextProtocolServers\": [ { \"transport\": { \"type\": \"stdio\", \"command\": \"npx\", \"args\": [\"-y\", \"mindbase-mcp\"] } } ] } }\n```\n\nMCP tools appear under `@` in chat.\n\n\n\n\nAny other MCP client (Zed, Aider, Goose, Claude Agent SDK…)\n\nPoint it at `npx -y mindbase-mcp` as a stdio server. See your client's MCP docs for the config location.\n\n\n\n**Next:** create your first project and learn the four daily workflows in the **[Guide →](docs/guide.md)**\n\n## The web UI\n\nSince 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/)**.\n\n
\n \n
\n\n**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.\n\n\nFeature matrix by editor\n\n| Feature | Claude Code | Cursor / Windsurf / Cline / Continue | Web UI |\n|---|---|---|---|\n| Slash commands (`/mb:*`) | ✅ | ❌ (use natural language) | ✅ (`/contribute`, `/build`, `/lint`, `/research`) |\n| MCP tools directly | ✅ | ✅ | ❌ |\n| Karpathy 8-step ingest with approval | ✅ (sub-agents) | ⚠️ Manual via prompt | ✅ (approval cards) |\n| Contribute / ingest PDF & URL | ✅ | ✅ | ✅ (upload + ✨ Process) |\n| Ask wiki with cited answers | ✅ | ✅ | ✅ |\n| Build / health check | ✅ | ✅ | ✅ |\n| Wiki tree browsing + rich editor | ❌ | ❌ | ✅ |\n| `-p project-id` routing | ✅ | ⚠️ natural language | ✅ (switcher) |\n\n\n\n## Where your data lives\n\nEverything is plain markdown under `~/mindbase-data/` (override: `MINDBASE_DATA_DIR`):\n\n```\n~/mindbase-data/projects/my-research/\n├── README.md # Ops manual — you edit, LLM reads\n├── context.md # Synthesized truth — LLM writes, you approve\n├── index.yaml # Auto-generated catalog\n├── sources/\n│ ├── contributors// # Your dated entries + notes (append-only)\n│ ├── research/ # LLM-authored wiki pages\n│ └── raw/ # PDFs, HTML captures\n├── logs/ # Chronological operation log\n├── artifacts/ # Briefs, exports, lint findings\n└── state/builder/snapshots/ # context.md snapshots for rollback\n```\n\nNo 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`.\n\n## Architecture at a glance\n\n```\n Claude Code / Cursor / any MCP editor Web UI (npx mindbase-app)\n │ MCP · sub-agents with │ /commands ·\n │ per-agent tool allowlists │ approval cards\n └─────────────┬─────────────────────────┬──────┘\n ▼ ▼\n ┌──────────────────────────────────────────────┐\n │ One ops engine: gather context → single │\n │ constrained JSON completion → human │\n │ approval → whitelisted executors → log │\n └─────────────────────┬────────────────────────┘\n ▼\n ~/mindbase-data/ · plain markdown\n (LLM: Ollama local models or any cloud key)\n```\n\nMonorepo: `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).\n\n## Docs & help\n\n- **[Guide](docs/guide.md)** — first project, the four daily workflows, multi-project routing, troubleshooting\n- **[Website](https://frankchu91.github.io/mindbase-llm-wiki/)** — live demos\n- **[CHANGELOG](CHANGELOG.md)** · **[Issues](https://github.com/frankchu91/mindbase-llm-wiki/issues)** — I reply to every issue same-day during beta\n- **The idea:** [Karpathy's LLM Wiki gist](https://gist.github.com/karpathy/442a6bf555914893e9891c11519de94f)\n\n## Roadmap\n\n**Next:** browser extension for one-click page capture · audio input via Whisper · unified meta-tool for Cursor/Windsurf slash-like UX.\n**Later:** team projects with human-in-the-loop review · audio digests · desktop app · mobile capture.\n\n## Feedback\n\nBeta 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.\n\n## License\n\n[MIT](LICENSE) — do what you want, no warranty. If you build something interesting on top, I'd love to hear about it.\n\n---\n\n**Built with the belief that AI's most valuable gift is not \"generation on demand\" but \"gardening of a persistent artifact you own.\"**\n",
"bytes": 11948,
"sha": "c278e610e00bf4e911d4d8297a145c2970473e31bcdb903cb9f24054078ee72a",
"repo_slug": "frankchu91/mindbase",
"fonte": "repo",
"truncated": false,
"api": "https://agentalog.com/api/listings/mcp_io_github_frankchu91_mindbase_eda0cdf6/readme"
}
![\"Demo:](\"docs/assets/contribute.gif\")
\n![\"MindBase:](\"docs/assets/hero-dark.svg\")
\n ![\"The](\"docs/assets/terminal-contribute.png\")
\n![\"MindBase](\"docs/assets/webui.png\")
\n