Skip to content
cyrylloPublic

About

Unofficial desktop app for Ollama. Local AI stack: multi-host Ollama, a LiteLLM model aggregator, and Open WebUI — no terminal, no Docker, no cloud.

Topics

Resources

Stars

0 stars

Watchers

0 watching

Forks

Repository files navigation

LamaDeck

A control panel for your local AI stack — a desktop app (PyQt6) for KDE. It started as a simple Ollama service manager and grew into one place for the whole stack: multi-host model management, a model aggregator (LiteLLM) that exposes one OpenAI-compatible endpoint for tools like Continue/VS Code, a browser chat UI (Open WebUI) and a built-in chat that can search the web, cite its sources and work with your files — also in your web browser, on your phone too. No terminal, no Docker, no cloud — everything runs on your LAN.

LamaDeck was called "Ollama Manager" until version 0.9. It's an independent project, not affiliated with Ollama. Updating from the old name keeps your settings and conversations — the app moves them on its first start.

(Polska wersja tego pliku: README_PL.md)

New here? The user guide shows every screen with screenshots — what each feature does and how to turn it on (also on the project page).

Status: 1.2 — stable. Daily driver on my own setup. New in 1.2: the chat in your web browser — the same chat, models and conversations, on your phone too — see the Roadmap. Bug reports and feedback are welcome. Project page: https://cyryllo.github.io/LamaDeck/

LamaDeck — Dashboard

Requirements

  • Python 3 + PyQt6, requests
  • systemd + polkit (pkexec) — standard on KDE/Debian; install.sh offers to install pkexec via apt if it's missing (needed for every action requiring admin rights)
  • curl — used by the app's own install buttons (Ollama, and uv for Open WebUI/LiteLLM); install.sh offers to install it via apt if it's missing
  • Ollama (if you don't have it, the app installs it with one click)
  • Optional: uv (for installing Open WebUI and LiteLLM — the app installs it itself if needed)
  • Optional: the QtPdf module for PDF attachments in the chat — included in the pip PyQt6; on Debian it's python3-pyqt6.qtpdf (install.sh offers it, the .deb pulls it in)
  • Optional: the QtWebEngine module (about 100 MB) for the "Page" view of the file preview in the chat — web pages and games shown inside the app. pip: PyQt6-WebEngine; Debian/Ubuntu: python3-pyqt6.qtwebengine (install.sh installs or offers it, the .deb and the AppImage include it). Without it, "Page" opens the file in your web browser
  • Optional: ffmpeg, pandoc, zstd (full Open WebUI functionality — voice, document RAG; the app offers to install these via apt right after a successful Open WebUI install)

Installation and running

Debian / Ubuntu: from the APT repository (recommended) — add it once, then LamaDeck updates together with the rest of the system:

sudo curl -fsSLo /usr/share/keyrings/lamadeck-archive-keyring.gpg \
  https://cyryllo.github.io/LamaDeck/apt/lamadeck-archive-keyring.gpg
sudo curl -fsSLo /etc/apt/sources.list.d/lamadeck.sources \
  https://cyryllo.github.io/LamaDeck/apt/lamadeck.sources
sudo apt update
sudo apt install lamadeck

The .deb from the releases page works too (sudo apt install ./lamadeck_*_all.deb) — it adds the same repository, so updates arrive the same way.

Any distribution: AppImage — one file with everything inside (Fedora, Arch, openSUSE, Ubuntu…). Download LamaDeck-<version>-x86_64.AppImage from the releases page, then:

chmod +x LamaDeck-*-x86_64.AppImage
./LamaDeck-*-x86_64.AppImage

On the first start it offers to add itself to the applications menu (also a switch in Settings); the menu entry and autostart follow the file if you move it. Needs glibc 2.34 or newer (Ubuntu 22.04, Debian 12, Fedora 35 and later).

From the source code — clone the repository and enter its directory — the install scripts below expect to be run from there (they look for lamadeck.py next to themselves):

git clone https://github.com/cyryllo/LamaDeck.git
cd LamaDeck

As a menu app, no root (recommended) — installs dependencies via pip, copies the app to ~/.local/share/lamadeck and adds a menu entry (Utilities section). Detects an existing install and offers to update/reinstall accordingly:

./install.sh

Uninstall with ./install.sh --uninstall.

Build the packages yourself — .deb (dependencies come from apt) or the AppImage:

./build-deb.sh && sudo apt install ./lamadeck_*_all.deb
./build-appimage.sh

Manually, for development — no copying, no menu entry:

pip install "PyQt6<6.12" "PyQt6-Qt6<6.12" "PyQt6-WebEngine<6.12" "PyQt6-WebEngine-Qt6<6.12" requests
python3 lamadeck.py

Features

The window has a sidebar with eight screens, in a dark Breeze look (Inter + JetBrains Mono fonts are bundled, so it looks the same everywhere).

Dashboard

  • One switch for all services (Ollama, LiteLLM router, Open WebUI) plus "Restart" and "Stop all"
  • A card per service with its status and its own on/off switch — or an "Install" button if it's missing (Ollama via its official script, the others via uv, no Docker, no admin rights)
  • Ollama card: which hosts are online, which models are loaded in memory, live GPU load
  • Router card: the routes (alias → model @ hosts) at a glance
  • Open WebUI card: one big button — install / start / open in the browser
  • Computer resources: CPU, VRAM and RAM gauges, network and CPU charts

Ollama hosts

  • All Ollama hosts (this computer and any on the LAN, e.g. BC-250): online/offline, Ollama version, response time, VRAM and RAM used by models (with the totals for this computer) and the models on each host — on disk or loaded into memory
  • On computers with an APU, the RAM figure includes the part of a model kept in system memory shared with the GPU (GTT) — shown as "Ollama X GB + APU Y GB"
  • Add a host by just typing its address (http:// and port 11434 are added for you), remove a LAN host with one click (this computer always stays); the list is shared by the router and the Models screen
  • For this computer: restart the Ollama service and jump to its log
  • A host that doesn't respond gets a hint on what to check (OLLAMA_HOST=0.0.0.0)

LiteLLM router

  • One OpenAI-compatible address (http://localhost:4000/v1) for all programs, combining the models of every host
  • A live route diagram: programs → LiteLLM → aliases → hosts
  • Aliases — fixed names such as chat or code for a model on chosen hosts; programs use the alias, so you can swap the model behind it without touching them
  • What to do when a model is on several hosts: take turns, least busy, or fastest
  • Choose which models the router offers under their own names (checkbox per model and host)
  • Optional API key — off by default (anyone on your network can use the models)
  • Ready-made settings for Continue.dev (a config.yaml to copy) and step-by-step instructions for connecting Open WebUI
  • Changes wait for "Restart the router" — no surprise downtime

Models

  • One search for two sources — the official Ollama library and Hugging Face (GGUF only); results unfold under the field with every variant and its real file size, plus a memory estimate. Type an exact name (llama3.2:3b, hf.co/...:Q4_K_M) and press Enter to download right away; pick the target host
  • "Only MoE models" checkbox — shows only mixture-of-experts models (and only their MoE variants), which run much faster on APUs and computers without a strong graphics card. Detected by description, name, Hugging Face tags and Ollama variant tags (e.g. 30b-a3b)
  • "Community models" checkbox — also shows models other authors published on ollama.com (ollama.com/author/model, e.g. jobautomation/OpenEuroLLM-Polish); off by default, because Ollama doesn't review them. Pasting an ollama.com link opens the model's variants directly
  • Download progress with size and speed, and a "Cancel" button (resumes later where it stopped); a desktop notification when a long download (or an install/update) finishes in the background
  • Tiles for the models on all hosts, filterable by kind (chat, code, embedding, vision): size, quantization, context length and on which hosts the model is (green = loaded in memory)
  • "Load" / "Unload" a model to/from memory, "Copy to host" (downloads the same model on another host), delete from one or all hosts
  • Hugging Face is a community source — the app suggests well-known quantization authors

Chat

Chat with speed statistics

  • A simple built-in chat with any model on any host, straight from Ollama — works without Open WebUI and without the router
  • Answers stream in as they're written, with formatting (code blocks, lists, tables) and a copy icon above every code block; stop at any time
  • Small icons in the message headers: copy an answer, regenerate the last answer (with the model selected right now — handy for comparing models) and edit the last question
  • While a model is being loaded into memory (the first question can take a while), the answer bubble says so and counts the seconds
  • Speed stats for every answer, measured by Ollama itself: generation and prompt processing in tokens/s, token counts, model loading time, total time — handy for comparing models and Ollama settings
  • Context usage: how much of the model's context window the conversation takes (e.g. "4,051 / 65,536 tokens"), with a warning from 80% — before the model silently starts forgetting the beginning
  • Thinking on/off for models that reason before answering (qwen3, gemma4…): off, a simple question takes a fraction of a second instead of several seconds; on, the reasoning can be shown in the conversation
  • Attach pictures, documents and text files: 📎 button, drag & drop or paste a screenshot (Ctrl+V). Pictures go to models that can see them (vision); text files (code, .txt, .md, .csv…) are added to the message
  • Documents: PDF, Word (.docx), Excel (.xlsx), PowerPoint (.pptx), LibreOffice (.odt, .ods, .odp) and EPUB — their text is extracted with nothing beyond Python and Qt (tables and sheets as a; b; c rows, slides and PDF pages numbered). Scanned PDFs without a text layer and old .doc/.xls/.ppt files are not supported (the chat says what to do)
  • Large text files (100 KB – 20 MB) are searched, not pasted: the file is split into excerpts and indexed with an embedding model in Ollama (e.g. nomic-embed-text); each question gets the best-matching excerpts, and the answer shows which lines they came from. Nothing beyond Ollama is needed, and nothing is stored on disk. The embedding model is chosen in Settings ("Chat" card); without one, a button there downloads the recommended multilingual nomic-embed-text-v2-moe
  • Web search with sources: the globe button next to the paperclip (shown once a results source is chosen in Settings) turns it on. The model writes 1–3 search queries, the app searches, reads up to 5 pages, picks the passages that match the question (with the embedding model) and the answer cites them as [1], [2]… with clickable sources below. Works with any model, no tool calling needed. Results come from SearXNG — installed with one click on this computer, no account — or your own SearXNG server, or Ollama's web search (free ollama.com account). Off by default: queries go to the internet only when you switch it on
  • Working folder: the folder button lets the model browse, read and search the files of a folder you choose and change them — every write shows a preview of the change and needs your approval (once, for the session, or always for that folder). The model can't leave the folder, delete files or run commands. Needs a model tagged tools (qwen3, gemma4…). The files it saved are listed under the answer; click one to open a preview next to the conversation — the code with line numbers, the changes, or the page itself (a web page or a game, as in a browser) — that refreshes by itself whenever the model saves the file again
  • Links in a question: paste a link and the app downloads the page itself — straight from your computer, no account — and adds its text to the question (a PDF link works too; long pages go through excerpt search). Pages that only load with JavaScript are reported as such
  • Conversations are saved automatically after every answer, one JSON file each in ~/.local/share/LamaDeck/czaty/ — only on your computer, untouched by updates and uninstalling. A list on the left opens, searches (titles and content), renames, exports to Markdown (.md) and deletes them; pictures are saved too, large files have to be attached again to keep searching them
  • Also opens straight from the tray panel ("Open chat")

Working folder: the model adds a button to index.html

Chat in the browser (1.2)

  • The same chat in your web browser — the same models and the same saved conversations — at http://localhost:11500; start at the computer, carry on from your phone. Answers as they're written, web search, links, attachments (paperclip, drag and drop, pasted screenshots), the working folder with the approval window, saved files and the file preview
  • Settings → Chat in the browser: on/off, port, access from the home network, the address and a QR code for your phone. Open in the browser on the Chat screen opens the current conversation
  • Off by default and only on this computer; the home network needs a password (stored as a hash, limited sign-in attempts). In the browser you can only use working folders chosen in the app; a page from the folder runs in a sandbox. While an answer is being written in one place, the other only reads that conversation — nothing gets overwritten
  • No new dependencies: a small web server inside the app, a plain HTML/JS page; Markdown via the bundled marked, the QR code via the bundled qrcodegen (both MIT)

Chat in the browser

Open WebUI

  • Start, restart, stop, open in the browser, copy the address
  • Settings: open the browser after start, require login (only switchable off while there are no accounts — otherwise Open WebUI would lock existing users out), port
  • Installation details: environment, where chats are stored, whether it's connected to the router; one-click backup of all chat data (a ZIP file)
  • After a fresh install, offers the optional ffmpeg/pandoc/zstd packages (voice, document RAG — the only step that needs admin rights)

Log

  • Messages from Ollama, LiteLLM, Open WebUI and the app itself in one place, filterable per service
  • Errors come with a plain explanation of what they mean and, where it applies, a "Show host" link
  • Save to a file, clear the view, auto-scroll
  • The Ollama service log needs permission: load it once (password) or allow it permanently

System tray (optional, off by default — turn it on in Settings)

Tray panel

  • A tray icon with a small panel: the installed services with on/off switches, CPU / VRAM / RAM, models in memory, and buttons to open the built-in chat or the control panel (plus a small "WebUI" button while Open WebUI is running)
  • Closing the window hides it in the tray; quit from the icon's menu
  • Optionally start the program after login, already hidden in the tray
  • Only one copy runs at a time — starting it again just brings the window back

Settings

  • Interface language — Polish, English, German, Spanish, French, Portuguese, Italian; switches instantly, remembered between runs
  • Autostart of all three services in one place: Ollama with the system, the router and Open WebUI after login
  • Versions and on-demand update checks for Ollama, LiteLLM, Open WebUI and SearXNG
  • Chat: the embedding model used to search large files (or a one-click download of the recommended one) and the web search source: off, SearXNG on this computer (install, start, stop, autostart, update, remove — for the user, no root), SearXNG at another address, or Ollama's web search with an API key; "Test search" runs one query
  • Services that aren't installed are hidden here, in the sidebar and in the tray panel
  • Ollama environment variables, each with a short description: OLLAMA_KEEP_ALIVE, OLLAMA_CONTEXT_LENGTH (default 4096 is too small for agentic work), OLLAMA_MAX_LOADED_MODELS, OLLAMA_NUM_PARALLEL, OLLAMA_FLASH_ATTENTION, OLLAMA_KV_CACHE_TYPE, OLLAMA_VULKAN, OLLAMA_IGPU_ENABLE, OLLAMA_HOST, IPAddressAllow (limit LAN access), GGML_VK_VISIBLE_DEVICES, OLLAMA_GPU_OVERHEAD, plus free-form extra variables — one "Save" applies everything with a single service restart

Roadmap

Done:

  • Multi-host model management, LiteLLM router, Open WebUI with one click, log with plain explanations, tray panel, seven languages
  • Built-in chat: attachments and documents, large files searched by excerpts, links in a question, saved conversations, context usage, thinking on/off, regenerate and edit, copy and Markdown export
  • 1.0 — web search in the chat with sources (SearXNG or Ollama's web search); own APT repository, .deb, AppImage
  • 1.1 — working folder in the chat: the model works with your files, every change approved
  • 1.1.x — default write permission per project, saved files under the answer, a file preview in the app that refreshes by itself (web pages and games too)
  • 1.2 — the chat in your web browser with the same features and conversations as in the window: answers as they're written, web search, attachments, working folder and file preview. Only this computer by default; the home network with a password and a QR code

Next: polishing from everyday use. Ideas and wishes are welcome in the issues.

Notes

  • Ollama service control always targets the local machine — even if a remote server (e.g. BC-250) is selected in the window, start/stop/autostart act locally.

About

Unofficial desktop app for Ollama. Local AI stack: multi-host Ollama, a LiteLLM model aggregator, and Open WebUI — no terminal, no Docker, no cloud.

Topics

Resources

Stars

0 stars

Watchers

0 watching

Forks

Releases

Contributors

Languages