Hardware Requirements
LanJAM runs AI models directly on your own computer, which means your hardware matters. The good news: most modern computers can run at least a basic AI model. Here's what you need to know.
The basics
To run LanJAM you'll need:
- 8 GB of RAM (memory) at a minimum — 16 GB or more is recommended
- 10 GB of free disk space — AI models are large files, so more space means you can try more models
- A reasonably modern operating system — macOS 13+, Windows 10/11, or a recent Linux distribution
LanJAM itself is lightweight. The real question is how much power you need to run AI models comfortably.
Apple Silicon Macs (M1, M2, M3, M4)
If you have a Mac with Apple Silicon, you're in luck. These chips share memory between the processor and the AI model, which means they can run surprisingly capable models even on entry-level machines.
| Your Mac's memory | What you can run |
|---|---|
| 8 GB | Small models (e.g. Llama 3.2 3B) — good for quick answers and simple tasks |
| 16 GB | Medium models (e.g. Llama 3.1 8B) — great for most everyday use |
| 32 GB | Large models (e.g. Llama 3.1 70B at reduced quality) — more detailed, nuanced responses |
| 64 GB+ | The largest openly available models with full quality |
Even an 8 GB MacBook Air can run a small AI model. Response speeds depend on your specific chip — newer chips (M3, M4) are faster than earlier ones.
Windows and Linux
On Windows and Linux, your experience depends on two things: your RAM and whether you have a dedicated graphics card (GPU).
- RAM: Same as Mac — 8 GB minimum, 16 GB recommended. The AI model needs to fit in memory.
- GPU (graphics card): If you have an NVIDIA graphics card with at least 6 GB of video memory, Ollama can use it to run models much faster. AMD GPUs have experimental support.
- SSD: An SSD (solid-state drive) is strongly recommended. AI models are large files and load much faster from an SSD than a traditional hard drive.
Do I need a GPU?
No. Ollama can run models using just your CPU — it'll just be slower. A dedicated NVIDIA GPU makes responses significantly faster, especially for larger models, but it's not required to get started.
If you're buying hardware specifically for LanJAM, an NVIDIA RTX 3060 (12 GB) or newer is a great choice. But try it on what you have first — you might be pleasantly surprised.
Running Ollama on a separate machine
If you have a powerful desktop or server but want to use LanJAM from other devices, you can run Ollama on one machine and LanJAM on another. During installation, the installer asks for the Ollama host address — just point it to the machine running Ollama (e.g. http://192.168.1.50:11434).
This is completely optional and an advanced setup. Most families run everything on a single computer.
Software you'll need
LanJAM requires two free tools:
- Docker — runs LanJAM's services (the web app, database, etc.) in isolated containers. Think of it as a tidy box that keeps everything contained. Available for macOS, Windows, and Linux.
- Ollama — runs AI models on your hardware. This is what makes everything private — your conversations stay on your network.
Both are free, open-source, and straightforward to install. The downloads page walks you through installing them step by step.
Still not sure?
If your computer was made in the last five years and has at least 8 GB of RAM, give it a try. The LanJAM installer will check your setup and let you know if anything is missing. You can always remove everything cleanly if it doesn't work out.