Saegor/Georges
1
1
Fork
You've already forked Georges
0
A minimalist CLI AI assistant.
  • Python 100%
2026年02月20日 13:55:03 +01:00
assets Mise à jour de la démo. 2026年02月10日 10:40:09 +01:00
georges Modification du design. 2026年02月20日 13:55:03 +01:00
LICENSE Ajout de la licence GPL 3.0 2026年01月05日 07:38:18 +01:00
main.py Ajout d'un contrôleur pour décharger le main.py 2026年01月23日 23:37:48 +01:00
README.md Modification mineure. 2026年02月13日 17:39:26 +01:00
requirements.txt Minimalisation. Transition vers la configuration dynamique. 2026年01月22日 07:48:41 +01:00

Georges
A Minimalist CLI AI Assistant

Yet Another Simple Interface for LLM Interaction

  • Developed on Linux but works anywhere if dependencies are met.
  • Used via terminal, it lets you chat with a remote LLM via API.
  • Easily save code snippets with a simple clipboard mechanism.

Attention! This is not vibe coding. If you want something like vibe coding, head to another project (like Toad, OpenCode, etc...)

Dependencies

# Debian/Ubuntu
apt install python3-setproctitle python3-prompt-toolkit python3-requests python3-markdown-it python3-pyperclip python3-rich
# Fedora/RHEL
dnf install python3-setproctitle python3-prompt-toolkit python3-requests python3-markdown-it python3-pyperclip python3-rich
# Arch Linux
pacman -S python-setproctitle python-prompt-toolkit python-requests python-markdown-it python-pyperclip python-rich
# pip
pip install -r requirements.txt

Configure

In georges/config.py, set PROVIDER, MODEL, MAX_TOKENS, TEMPERATURE, and SYSTEM_MESSAGE.

# EXAMPLE
SELECT = "MIST-CODE"
CONFIGS = {
 "MIST-CODE": {
 "PROVIDER": "MISTRAL",
 "MODEL": "codestral-latest",
 "MAX_TOKENS": 1024,
 "TEMPERATURE": 0.2,
 "SYSTEM_MESSAGE": "Minimalist code. Short answers",
 },
 # Add more configs as needed
}

API KEYS

Option 1: Environment Variables

# Unix
export MISTRAL_API_KEY="..."
export ANTHROPIC_API_KEY="sk-ant-..."

Option 2: .env File

Create a .env file in the main folder:

# .env
MISTRAL_API_KEY=...
ANTHROPIC_API_KEY=sk-ant-...

Run

python main.py

Usage

  • Simple Input: Type your message and press ENTER.
  • Pager: Long responses open in a pager (q to quit).
  • Multiline Input:
    • Send m to enable multiline mode.
    • Submit with Alt+Enter or Esc Enter.
  • Help: Send h to see the available commands.
  • Config: Send c to see your LLM preset.
  • Token usage : Send t to see your token usage.
  • Code: Send e to extract code.
  • Reset Session: Send r to clear history and token count.
  • Change Preset: Send l to load a new preset.
  • Exit: Send empty input (ENTER) to end the chat.

Screenshot

Screenshot

Demo

Demo