Now since AI is a hype we made a light shell for run your ollama in a browser.
There are nice tools like open webui that can do the same thing but something
that you can customize for your needs are better and have all the basic
functions.
The cell is light and you can custom to serve your needs.
You need a html file for the shell and one python for the use of Ram and Vram.
The files you can download it and unzip it.
Works in Linux,Windows(haven't checked) and maybe Apple Mac.
Files click here to download the files.
Required
| Package | Purpose | Likely already installed if... |
|---|---|---|
| Ollama | Runs the LLM models | you need Ollama |
| Python 3 | Runs sysmon.py | virtually every Linux distro ships this by default |
| NVIDIA driver (provides nvidia-smi) | VRAM stats + GPU temp | you're using an NVIDIA GPU for Ollama — already confirmed working |
| bash | Runs start.sh | default shell interpreter on all major distros, even if you use zsh interactively |
| curl | Used inside start.sh to check if sysmon is running | usually preinstalled; if not: |
| A browser: Chrome or Edge | Runs the actual chat page | you're already using Chrome |
Install commands by distro (if anything's missing)
Debian/Ubuntu:
sudo apt install python3 curlFedora:
sudo dnf install python3 curlArch:
sudo pacman -S python curlOllama and the NVIDIA driver are installed separately from their own official sources, not your distro's default repos:
- Ollama:
curl -fsSL https://ollama.com/install.sh | sh - NVIDIA driver: via your distro's driver manager, or NVIDIA's .run installer
Nothing else needed
- No
pip installpackages — sysmon.py is 100% Python standard library - No
npm install— the HTML page has zero build step, just open it in the browser