GGUF Model Manager for llama-swap — browse HuggingFace, download, manage models, all in one place.
A lightweight web UI to manage your local GGUF model collection.
llama-swap.yaml with automatic reloadTested on:
Should work, but untested — same apt base, no reports either way:
Non-apt distros (Fedora, Arch, openSUSE): system dependencies must be installed manually.
git clone https://github.com/miradorventus/EZmodL.git
cd EZmodL
EZMODL_LOCAL_INSTALL=1 bash install.sh
Or as a one-liner:
curl -fsSL https://raw.githubusercontent.com/miradorventus/EZmodL/main/install.sh | bash
📌 The installer downloads the files, installs system dependencies (via pkexec), sets up the menu and desktop entries, and tells you if anything is missing.
EZmodL is designed to work with llama-swap, so you need:
llama-swap configured (~/.llamaui/config/llama-swap.yaml)llama.cpp built with your backend (Vulkan / CUDA / ROCm)For a fully automated llama.cpp + llama-swap setup on AMD, see ollama-amd-plug-and-play.
Qwen3) or paste a repo (bartowski/Qwen_Qwen3-14B-GGUF)All configuration lives in ~/.ezmodl/:
~/.ezmodl/
├── config/
│ ├── preferences.json # VRAM target, sort order, curators_only, theme
│ └── curators.txt # Editable curator list
├── queue/ # Download queue and history
└── logs/ # Activity log
Editing the curator list (add or remove):
nano ~/.ezmodl/config/curators.txt
The CSS is exposed at ~/.ezmodl/static/style.css. CSS variables sit at the top of the file:
:root[data-theme="dark-futurist"] {
--bg-base: #0a0e1a;
--text-accent: #00d9ff;
/* ... */
}
You can build your own theme by adding a :root[data-theme="my-theme"] block.
| Component | Role |
|---|---|
server.py |
Flask backend, exposes a REST API |
lib/hf_api.py |
HuggingFace API wrapper (search + tree) |
lib/vram_classify.py |
AMD/NVIDIA VRAM detection and quantization classification |
lib/disk_search.py |
Multi-folder search and symlink creation |
lib/yaml_helper.py |
Add/remove models in llama-swap.yaml |
lib/parse_quant.py |
Extracts the quantization from a filename |
templates/index.html |
Main UI (collapsible sections) |
static/app.js |
Vanilla JS, fetches the REST endpoints |
ezmodl.sh |
Launcher, OllamaUI pattern (lockfile, integrity check) |
cd EZmodL
bash uninstall.sh
GGUF models in ~/llm-models/ are kept. llama-swap.yaml is left untouched — clean it up yourself if needed.
Issues and PRs welcome. This project is deliberately simple and minimal — let’s keep it that way.
MIT — see LICENSE.
EZmodL — because managing GGUFs shouldn’t be a chore.