Why?

I created this out of frustration. I always work around Neovim and tmux, and whenever I want advice or have a question in mind, I use AI any AI. But that usually forces me to use a mouse, which I don’t want to do because I want to work in the terminal as much as possible.

That’s when I decided to just vibe-code this thing so I can have the app asap, and also to experience what it feels like to vibe-code (I hate it, btw xd). But after discovering the free API from NVIDIA NIM, that really triggered this idea.

Even though this project was just vibe-coded, I still have the intention to continue it as a hobby. I’ll probably:

• Make it modular — everything is in a single file right now, which is ugly.
• Add history? — Initially, I didn’t want to because I rarely use it and I always copy the response right after it’s generated.
• Change the model — NVIDIA NIM has a variety of LLMs, and it might be better to switch to a faster one since it’s currently using R1.

Anyway, everything here is AI-generated except for this section.

Features ✨

• 🗣️ Interactive chat interface with streaming responses
• 📎 File attachment support (PDF, DOCX, TXT, code files, etc.)
• 📋 Clipboard integration for responses
• 🔍 Context-aware conversations
• 📁 Home directory-relative file paths
• 🚦 Progress indicators and error handling

Quick Start 🚀

# Install dependencies
pip install -r requirements.txt

# Set up your .env file
echo "NVIDIA_API_KEY=your_api_key_here" > .env

# Run the application
python app.py

Installation 📦

• Ensure Python 3.8+ is installed
• Clone this repository
• Install the app:

  ```bash
   pip install .
  

• Create .env file with your NVIDIA API key

Basic Usage 💻

terminal-nim
> attach
> /path/to/your/file.pdf
> done
> Explain the attached document

Supported File Types 📄

➡️ Full Documentation