Skip to main content

An open-source, local AI chatbot built with GTK4

Project description

Chi

GitHub Repo stars GitHub release (latest by date) GitHub license PyPI status

Install

1. Install system dependencies

Chi uses GTK4 for its interface. These cannot be installed via pip and must come from your system package manager.

Debian / Ubuntu

sudo apt install python3-gi python3-gi-cairo gir1.2-gtk-4.0 libgtk-4-1

Fedora

sudo dnf install python3-gobject gtk4

Arch

sudo pacman -S python-gobject gtk4

2. Install Chi

It is recommended to use pipx to install Chi, as it keeps Chi and its dependencies isolated from your system Python.

pipx install chi-ai

Don't have pipx? Install it first:

pip install pipx

Or if you'd prefer plain pip:

pip install chi-ai

3. Run

chi-ai

Features

This section describes (and shows how to use) all features in Chi.

Model selection

Model selection is important, as Chi will require a specificed model to use in order to continue. Once selected, it will be saved in config.ini and you will not need to redo this process unless you are switching models.

In the settings menu of Chi, click "Open file", located underneath the text that says "Change your model. Upon clicking, a file dialog will appear, allowing you to look through your computer for a model to use. Note that it must be a .gguf model, optimized for LlaMA 3 style prompts.

After selecting a model, you will be able to use Chi in the "Chat" menu.

Generation parameters

In the settings menu of Chi, you will notice two sliders. One controls generation tokens, and the other controls generation temperature. The following is a quick description of these parameters.

  • Tokens are the maximum length of a response. The higher the number, the longer the response.
  • Temperature is the 'creativity' of the model. A lower number means more logical, basic responses, while a higher number means higher creativity, at the sacrifice of some logic.

Quick action buttons

In the chat section of Chi, you may notice three buttons. These buttons append a message at the end of an input, and can alter responses in more favorable ways. The translations to the prompt for each button is provided in the following.

  • Explain like I'm 5 -> Explain this in a way suitable for a 5 year-old.
  • Explain thoroughly -> Explain this thoroughly with examples.
  • Explain efficiently -> Explain this efficiently and concisely.

Personality customization

In the settings section of Chi, there is an entry box to change the personality. You can write instructions and press enter to save them and apply them to the model. Your customizations are saved in config.ini so there is no worry regarding loss of settings.

Want to contribute?

Read CONTRIBUTING.md for information regarding contribution. Any issues can go into the 'issues' section of the repository.

License

This project is licensed under the MIT License. See the LICENSE file for details.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

chi_ai-0.6.2.1.tar.gz (12.2 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

chi_ai-0.6.2.1-py3-none-any.whl (11.0 kB view details)

Uploaded Python 3

File details

Details for the file chi_ai-0.6.2.1.tar.gz.

File metadata

  • Download URL: chi_ai-0.6.2.1.tar.gz
  • Upload date:
  • Size: 12.2 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.2.0 CPython/3.13.5

File hashes

Hashes for chi_ai-0.6.2.1.tar.gz
Algorithm Hash digest
SHA256 966f75f83246b95effe7c6db969316e5b845ca51eba9ca8546b65b056e864b48
MD5 f47cc96aeb9d0858ce6b8b3779f55a77
BLAKE2b-256 1451f8dc32f4180099f7b5fb8d20033559a9ee4c58a70a6ea4ceffa40ce150b9

See more details on using hashes here.

File details

Details for the file chi_ai-0.6.2.1-py3-none-any.whl.

File metadata

  • Download URL: chi_ai-0.6.2.1-py3-none-any.whl
  • Upload date:
  • Size: 11.0 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.2.0 CPython/3.13.5

File hashes

Hashes for chi_ai-0.6.2.1-py3-none-any.whl
Algorithm Hash digest
SHA256 fcd7adf6d07239c83b523739a12baf081a561730b2800ce5f2bd2396058314e8
MD5 5a03e068cd7d6534cfeb54cd751b7a6d
BLAKE2b-256 56dfc3f9ccd3e2cecf28d3d4c8319c1643a9466b6d4df02b72fcc5a514f1f817

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page