TerminalAI

TerminalAI is a simple terminal emulator that integrates local LLMs via Ollama to provide intelligent command-line assistance using /help. Use your terminal as normal — and get smart suggestions right when you need them.

---

✨ Features

• Embedded terminal interface with ImGui
• Local LLM integration via Ollama (e.g., deepseek-coder, phi3, llama3)
• Instantly get shell command suggestions with /help
• Offline-first: No need for OpenAI API keys
• Configurable backend LLM

---

📦 Installation

You can either compile from source or download a prebuilt binary.

---

🔧 Compile from Source

1. Clone the Repository

git clone https://github.com/nigelloh15/termai.git
cd terminalai

2. Configure Python Version

• Open CMakeLists.txt
• Update the Python version reference to match the Python version installed on your system (e.g., Python3.12 or Python3.11)

3. Set Preferred LLM

• Open integration.py
• Set the variable or hardcoded model name to the local LLM you want to use (e.g., "deepseek-coder" or "phi3")

4. Build the App

mkdir build
cd build
cmake ..
make

• This will generate a .app file for MacOS.

---

📁 Prebuilt Binary

Download the latest .app file from the Releases tab.

• The prebuilt version uses:

  • Python: 3.12.1
  • Default LLM: deepseek-coder via Ollama

No setup needed beyond ensuring ollama and the model are available on your machine.

---

🚀 Usage

1. Install Ollama

If you haven’t already, install Ollama:

curl -fsSL https://ollama.com/install.sh | sh

2. Pull a Model

Pull the model you'd like to use (must match the one specified in integration.py if building from source):

ollama run deepseek-coder
# or for other models:
ollama pull phi3

💡 Make sure the model is running or available before launching TerminalAI.

3. Launch TerminalAI

If Built Locally:

./build/TerminalAI.app  # macOS .app

If Prebuilt:

Double-click the .app file or run it from terminal.

---

💬 How to Use

• Type /help <your question> to ask for CLI assistance Example:

  /help how to list all hidden files
  

• You will get a suggestion from the local LLM.

• Use the terminal like any other shell.

---

🛠️ Configuration

You can swap out the LLM used by modifying integration.py and restarting the app. Make sure the model is installed via ollama.

---