40.6K ·GitHub
What is this?
Project NOMAD is a self-contained, offline-first knowledge and education server: offline Wikipedia, thousands of books, Khan Academy courses, offline maps and optional local AI, all running on hardware you own with no internet required.
Project NOMAD is a self-contained, offline-first knowledge and education server: offline Wikipedia, thousands of books, Khan Academy courses, offline maps and optional local AI, all running on hardware you own with no internet required.Is it one app or a bundle of separate tools?It is a management UI (the Command Center) plus an API that orchestrates a stack of containerized tools via Docker — Kiwix, Kolibri, ProtoMaps, CyberChef, FlatNotes, Ollama — handling installation, configuration and updates for everything.
It is a management UI (the Command Center) plus an API that orchestrates a stack of containerized tools via Docker — Kiwix, Kolibri, ProtoMaps, CyberChef, FlatNotes, Ollama — handling installation, configuration and updates for everything.How do I get it running?Run the terminal install script on any Debian-based OS (Ubuntu 26.04 LTS recommended; Windows users have a community WSL2 guide), then open a browser. Advanced users can deploy the Docker Compose template with docker compose up -d instead.
Run the terminal install script on any Debian-based OS (Ubuntu 26.04 LTS recommended; Windows users have a community WSL2 guide), then open a browser. Advanced users can deploy the Docker Compose template with docker compose up -d instead.What kind of hardware do I need for this?The core is light: a 2 GHz dual-core CPU, 4 GB RAM and about 5 GB of disk. The AI Assistant is what demands muscle — an RTX 3060-class GPU, 32 GB RAM and 250 GB disk, since a model needs roughly its download size in memory while answering.
The core is light: a 2 GHz dual-core CPU, 4 GB RAM and about 5 GB of disk. The AI Assistant is what demands muscle — an RTX 3060-class GPU, 32 GB RAM and 250 GB disk, since a model needs roughly its download size in memory while answering.Can I safely put it on my home network for other devices?It ships with no authentication by design and is not meant to face the internet. On a local network, control what other devices can reach by blocking or opening ports — direct internet exposure is strongly advised against.
It ships with no authentication by design and is not meant to face the internet. On a local network, control what other devices can reach by blocking or opening ports — direct internet exposure is strongly advised against.Who would actually use something like this?Anyone who wants a full knowledge stack without internet: the project positions itself against bare-minimum offline survival kits, and Kolibri's multi-user support plus planned parental and teacher accounts point at family and classroom use.
Anyone who wants a full knowledge stack without internet: the project positions itself against bare-minimum offline survival kits, and Kolibri's multi-user support plus planned parental and teacher accounts point at family and classroom use.Do I need internet after the initial install?No — the connection is only needed during installation to download dependencies, and later only if you choose to fetch more content or updates. After that, everything runs fully disconnected.
No — the connection is only needed during installation to download dependencies, and later only if you choose to fetch more content or updates. After that, everything runs fully disconnected.Does it collect any telemetry about me?No — zero built-in telemetry. The only outbound check is a configurable connectivity test (Cloudflare's 1.1.1.1 by default, with fallbacks) so the UI can show whether you are online.
No — zero built-in telemetry. The only outbound check is a configurable connectivity test (Cloudflare's 1.1.1.1 by default, with fallbacks) so the UI can show whether you are online.What exactly comes bundled out of the box?Offline Wikipedia, medical references and e-books via Kiwix, Khan Academy courses with progress tracking via Kolibri, offline regional maps via ProtoMaps, CyberChef data tools, FlatNotes markdown notes, a hardware benchmark and the Supply Depot app catalog.
Offline Wikipedia, medical references and e-books via Kiwix, Khan Academy courses with progress tracking via Kolibri, offline regional maps via ProtoMaps, CyberChef data tools, FlatNotes markdown notes, a hardware benchmark and the Supply Depot app catalog.Can I add my own apps or containers to it?Yes — the Supply Depot offers one-click installs (PDF tools, file browser, e-book library, password manager and more), and you can also run your own custom Docker containers alongside the built-in stack.
Yes — the Supply Depot offers one-click installs (PDF tools, file browser, e-book library, password manager and more), and you can also run your own custom Docker containers alongside the built-in stack.Does it update itself?Optionally: opt-in auto-updates install minor and patch versions of the core, installed apps and offline content on a schedule you set, after pre-flight checks. Major versions always require a manual update.
Optionally: opt-in auto-updates install minor and patch versions of the core, installed apps and offline content on a schedule you set, after pre-flight checks. Major versions always require a manual update.Can the AI assistant run on a different machine than the NOMAD server?Yes — in the AI Assistant settings you can point it at a remote Ollama or OpenAI-compatible server such as LM Studio. A remote Ollama must be started with OLLAMAHOST=0.0.0.0, and you handle that host's setup and model downloads yourself.
Yes — in the AI Assistant settings you can point it at a remote Ollama or OpenAI-compatible server such as LM Studio. A remote Ollama must be started with OLLAMAHOST=0.0.0.0, and you handle that host's setup and model downloads yourself.What are similar projects?QAnything, private-gpt, sim, Vane, delivery-harness, Companion-Space, DB-GPT, meetily
QAnything, private-gpt, sim, Vane, delivery-harness, Companion-Space, DB-GPT, meetilyWhich topics is it in?maps, homelab, offline-first, openai-compatible, local-llm, knowledge-base, ollama, rag, desktop, gpu, privacy, education
maps, homelab, offline-first, openai-compatible, local-llm, knowledge-base, ollama, rag, desktop, gpu, privacy, education