2024-05-12 18:20:17 -06:00
2024-05-16 10:01:08 -06:00
2024-05-16 11:37:35 -06:00
2024-05-12 18:18:25 -06:00
2024-05-14 00:27:02 -06:00
2024-05-15 23:05:06 -06:00

Alpaca

An Ollama client made with GTK4 and Adwaita.


Warning

This project is not affiliated at all with Ollama, I'm not responsible for any damages to your device or software caused by running code given by any models.

Important

This is my first GTK4 / Adwaita / Python app, so it might crash and some features are still under development, please report any errors if you can, thank you!

Features!

  • Talk to multiple models in the same conversation
  • Pull and delete models from the app

Future features!

  • Multiple conversations
  • Image / document recognition
  • Notifications
  • Code highlighting

Screenies

Login to Ollama instance Chatting with models Managing models
Screenshot from 2024-05-12 19-58-28 Screenshot from 2024-05-12 20-01-08 Screenshot from 2024-05-12 20-01-31

Preview

  1. Clone repo using Gnome Builder
  2. Press the run button

Instalation

  1. Go to the releases page
  2. Download the latest flatpak package
  3. Open it

Usage

  • You'll need an Ollama instance, I recommend using the Docker image
  • Once you open Alpaca it will ask you for a url, if you are using the same computer as the Ollama instance and didn't change the ports you can use the default url.
  • You might need a model, you can get one using the box icon at the top of the app, I recommend using phi3 because it is very lightweight but you can use whatever you want (I haven't actually tested all so your mileage may vary).
  • Then just start talking! you can mix different models, they all share the same conversation, it's really cool in my opinion.

About forks

If you want to fork this... I mean, I think it would be better if you start from scratch, my code isn't well documented at all, but if you really want to, please give me some credit, that's all I ask for... And maybe a donation (joke)

Description
An Ollama client made with GTK4 and Adwaita
Readme 52 MiB
Languages
Python 96.2%
Meson 2.6%
Shell 0.7%
CSS 0.5%