yap-shell

Your overly chatty terminal companion running Qwen 3.6 via Ollama!

Features

• Funny phrases — every interaction is a laugh
• Cool ASCII art — banners, cats, robots and more
• Animations — typing effects, spinners, rainbow text, sparkles
• Border-box panels — responses displayed in styled boxes
• Slash commands — /joke, /art, /fact, /status, etc.
• Conversation context — remembers the last 20 turns
• Zero deps — runs with just Python + Ollama installed

Requirements

• Python 3.9+
• Ollama with qwen3.6 model loaded

Install

pip install yapshell
yap

On first launch, yap runs a one-time setup that checks for Ollama, starts it,
and pulls the qwen3.6 model. Rerun it any time:

yap --setup            # interactive
yap --setup --yes      # no prompts
yap --setup --check    # just report status
yapshell-setup         # same thing, standalone

Or directly:

python3 yapshell/main.py

Commands

Command	Description
(anything)	Chat with YAP
/help	Show command list
/joke	Tell a joke
/art	Show ASCII art
/fact	Random fun fact
/clear	Clear the screen
/status	System stats
/compliment	Get complimented
/quit	Exit gracefully

How It Works

yap-shell talks to Ollama's CLI directly via subprocess — no Python API library needed. The qwen3.6 model must be pulled and running:

ollama pull qwen3.6
# (optional) ollama run qwen3.6  # keep this running, or yap will auto-start it
python3 main.py

About the AI:

By default, the AI is instructed to:

        "You are YAP, the overly chatty AI companion of yap-shell. "
        "Be funny, enthusiastic, and slightly unhinged but still helpful. "
        "Keep responses concise and engaging. "
        "If you don't know something, make up a funny response instead of saying you're an AI. "
        "Speak in first person like a weird friend who knows too much. "
        "Be creative and colorful with your language."