gknowledge

English · 简体中文 · Italiano

Karpathy's LLM Wiki as a native GNOME app.
Drop in your documents. Get a linked wiki that an LLM keeps up to date.
A native GNOME app. Offline by default, plain Markdown on disk, Obsidian-compatible.
Bring your own brain: a local model, any OpenAI-compatible endpoint, or your own ChatGPT plan.

_{Real output of the offline engine on six AI papers (data/demo/papers), no API key. Regenerate with make demo-svg.}

Install

One line, on Ubuntu 24.04+ and Debian 13+. It adds the signed gknowledge APT repository, so updates arrive with apt upgrade (other distributions get a per-user install):

curl -fsSL https://github.com/openapi/gknowledge/releases/latest/download/install.sh | bash

Or add the APT repository yourself. apt also pulls GTK 4, Libadwaita and PyGObject:

sudo curl -fsSLo /usr/share/keyrings/gknowledge.gpg https://openapi.github.io/gknowledge/gknowledge.gpg
sudo curl -fsSLo /etc/apt/sources.list.d/gknowledge.sources https://openapi.github.io/gknowledge/gknowledge.sources
sudo apt update && sudo apt install gknowledge

On any distribution, with Flatpak (the GNOME runtime comes from Flathub):

flatpak install --user https://openapi.github.io/gknowledge/io.gknowledge.Gknowledge.flatpakref

Both repositories are described at openapi.github.io/gknowledge.

Then open gknowledge from the app grid. To remove it: sudo apt remove gknowledge (or flatpak uninstall io.gknowledge.Gknowledge).

Other distributions, from source, and release builds

• Fedora, Arch and others: install PyGObject, GTK 4 and Libadwaita with your package manager, then run the one-line installer above. It installs into ~/.local without root.
• From source: git clone https://github.com/openapi/gknowledge && cd gknowledge && make install.
• Single files: every version's .deb and Flatpak bundle is on the Releases page, e.g. sudo apt install ./gknowledge_all.deb after downloading gknowledge_all.deb.

Why gknowledge

If you have used LLM Wiki, Tolaria, OpenKnowledge, Reor or Khoj, you already know the loop: ingest sources, let the model build and maintain a wiki, then ask questions against it. gknowledge does the same thing as a real GNOME application.

• Native, not a web view. GTK 4 and Libadwaita, built to follow the GNOME HIG. No Electron, no Tauri, no browser engine.
• Works with no account. A built-in heuristic engine ingests and answers offline. Add a model when you want better synthesis.
• Your data stays yours. Markdown with YAML frontmatter plus a readable SQLite index. Open the folder in Obsidian or any editor.
• Cited, reviewable, incremental. Every generated page records its sources, answers cite pages, and a review queue keeps you in control.

The workflow you already know

You know this from…	What it is	In gknowledge
Karpathy's LLM Wiki, LLM Wiki	Ingest: read a new source, update the pages it touches	Done: two-step extraction with provenance
Reor, Khoj	Ask: questions answered from your notes, with citations	Done: lexical search plus graph expansion, citations, saved answers
LLM Wiki, OpenKnowledge	index.md and log.md, [[wikilinks]], Obsidian-compatible vault	Done
LLM Wiki, Tolaria	Review AI changes before they stick	Review queue done, Git diff planned
Karpathy's LLM Wiki, OpenKnowledge	Lint: orphans, contradictions, stale pages	Planned
LLM Wiki	Knowledge graph, clusters, gaps	Planned
Tolaria, OpenKnowledge, LLM Wiki	MCP server so agents can read and write the wiki	Done: gknowledge-mcp
LLM Wiki	Web clipper, folder watch, deep research	Planned

See ROADMAP.md for dates and TODO.md for the backlog.

Karpathy's LLM Wiki, native on GNOME

In April 2026 Andrej Karpathy described the LLM Wiki pattern: instead of re-reading raw documents on every question, an LLM builds and maintains a persistent wiki that compounds over time. gknowledge implements it as a desktop app.

What we kept

• Three layers: immutable sources/, the generated knowledge/ wiki, and settings that tell the engine how to work.
• The operations: ingest (read a source, update the pages it touches) and query (answer from the wiki, with citations). Lint is next.
• index.md as the catalogue and log.md as the append-only history.
• Plain Markdown with [[wikilinks]] and YAML frontmatter, so the folder opens in Obsidian.
• The human curates, the model maintains.

What we changed and added

• A native app instead of an idea file. GTK 4 and Libadwaita, with sources, wiki, assistant and review in one window.
• Works with no model at all. An offline engine ingests and answers; a model makes it better.
• Bring your own brain. Your ChatGPT plan through Sign in with ChatGPT, Ollama, or any OpenAI-compatible endpoint.
• Two-step ingest with provenance. Analysis first, then pages; every page records its sources.
• No empty pages, no duplicates. Pages are created only with real content, names are merged across sources ("Kasparov" and "Garry Kasparov"), minor names stay as mentions until a second source cites them.
• A review queue for what the model is unsure about.
• An MCP server, so Claude Code, Codex or Cursor can search, read and grow the same wiki.

Use your wiki from Claude Code, Codex or Cursor

gknowledge ships an MCP server. Point any agent at a project folder:

claude mcp add gknowledge -- gknowledge-mcp ~/Knowledge/my-project     # Claude Code
codex mcp add gknowledge -- gknowledge-mcp ~/Knowledge/my-project      # Codex CLI

For Cursor or other clients, add this to the MCP configuration:

{ "mcpServers": { "gknowledge": { "command": "gknowledge-mcp", "args": ["/home/you/Knowledge/my-project"] } } }

Tool	What the agent can do
search_wiki	find pages by keyword
read_page	read a page with its sources and backlinks
list_pages	list pages by type
ask	get a cited answer, optionally saved back into the wiki
add_source	import and ingest a file
recent_activity	see what changed recently

Try it from source

sudo apt install python3-gi gir1.2-gtk-4.0 gir1.2-adw-1   # Ubuntu / Debian
git clone https://github.com/openapi/gknowledge && cd gknowledge
make start      # opens a demo project with six AI papers already ingested

make start also restarts the app whenever you edit the source, which makes it the fastest way to hack on gknowledge. Install it for daily use with make install.

Bring your own brain

Provider	Needs	Notes
Offline (default)	nothing	Heuristic extraction and extractive answers. Fully private.
ChatGPT plan	A Plus or Pro account	Preferences, Provider, Sign in with ChatGPT. Usage is charged to your own plan, no API key.
Ollama / LM Studio / vLLM	A local server	Any OpenAI-compatible endpoint on your machine.
OpenAI-compatible cloud	An API key	OpenAI, OpenRouter, Together, and similar.

Nothing is sent anywhere unless you pick a provider that is not offline.

How it compares

	gknowledge	LLM Wiki	Tolaria	OpenKnowledge	Reor (archived)	Khoj
Native GNOME / GTK 4	yes	Tauri	Tauri	Electron	Electron	web + Electron
Works offline, no key	yes	needs a model	AI optional	agent-driven	local model	self-host
Plain Markdown vault	yes	yes	yes	yes	yes	reads yours
Linux packaging	.deb, Flatpak planned	.deb, AppImage	AppImage, .deb	.deb, .rpm	AppImage	pip, Docker
License	MIT	GPL-3.0	AGPL-3.0	GPL-3.0	AGPL-3.0	AGPL-3.0

Details and history of each project are in DISTRIBUTING.md.

gknowledge is an alternative to…

gknowledge sits at the intersection of LLM-powered knowledge bases, local-first note tools, and document RAG systems. If you have used any of the tools below, you already understand the shape of the problem — gknowledge addresses it as a native GNOME/GTK application that is local-first, provider-agnostic, and usable fully offline (no API key required, thanks to a built-in heuristic extraction engine).

LLM wiki / auto-knowledge-base builders (the closest peers, several studied in NOTES.md):

• LLM Wiki (nashsu/llm_wiki) — Tauri desktop app implementing Karpathy's incremental "LLM builds and maintains a wiki" pattern
• Tolaria — files-first, git-first Markdown knowledge base with AI-assisted editing
• WeKnora (Tencent) — enterprise RAG + agents + auto-wiki framework
• OpenKnowledge (Inkeep) — WYSIWYG Markdown editor with agent/MCP integrations for LLM wikis
• Reor — local-first AI note-taking that auto-links related notes
• Khoj — self-hostable AI "second brain" over your documents

Not a peer: Google NotebookLM offers a similar cited-notebook experience but is closed-source and cloud-only.

Document RAG / "chat with your knowledge" engines:

• AnythingLLM, Onyx (formerly Danswer), Quivr, PrivateGPT, GPT4All, Cherry Studio, Verba (Weaviate)

Local-first / networked-thought note tools (which gknowledge complements with automated ingestion and provenance):

• Obsidian (+ Smart Connections / Copilot plugins), Logseq, SiYuan, Trilium Notes, Zettlr, Foam, Dendron, Joplin, Notion, Tana, Capacities, Heptabase, RemNote

Document / reference managers with growing AI features:

• Zotero, Paperless-ngx

How gknowledge differs

• Native to GNOME, not Electron/Tauri or web-first — GTK 4 + Libadwaita, HIG-aligned, Flatpak-friendly.
• Offline-first and provider-agnostic — a local heuristic engine ships in the box; OpenAI-compatible endpoints (OpenAI, Ollama, LM Studio, vLLM, OpenRouter, …) are optional, swappable, and configured per project.
• Inspectable, user-owned format — knowledge is plain Markdown with YAML frontmatter (Obsidian-compatible) plus a readable SQLite index; nothing is locked behind a sync service.
• Incremental, not stateless RAG — sources are ingested once into a persistent, cross-linked knowledge space with provenance, rather than re-retrieved from scratch on every question.
• Human-in-the-loop — generated content is clearly marked and a review queue keeps you in editorial control.

Project docs

• ROADMAP.md — milestones and the features we are borrowing from peers
• TODO.md — actionable backlog
• DISTRIBUTING.md — positioning and distribution study of comparable projects (Italian)
• CONTRIBUTING.md — how to contribute
• docs/DESIGN.md — the original product and design document
• CHANGELOG.md — what changed in each release

License

MIT. Contributions are welcome, see CONTRIBUTING.md.

Star history

If gknowledge is useful to you, a star helps other GNOME users find it.