🌌 Local Chatbot for LM Studio

Created by Arun Verma

Python Version: 3.10+ | Backend: LM Studio | UI: CustomTkinter | License: MIT

A sleek, standalone desktop client for running local Large Language Models (LLMs). Designed to connect seamlessly to the LM Studio local server, this application provides a robust, professional-grade interface with live hardware monitoring, document handling, and native support for modern reasoning models.

✨ Key Features

• 🎨 Windows 7 Aero Aesthetic: Features a clean, highly readable interface with classic "Aero Light" and "Dark Mode" skeuomorphic themes.

• 📊 Live Hardware Telemetry: Real-time percentage tracking for CPU, System RAM, and NVIDIA GPU VRAM utilization directly in the sidebar.

• 🧠 Deep Reasoning Support: Native parser for DeepSeek-R1 and similar models. Automatically intercepts <think> tags and displays the model's internal reasoning trace in a dedicated, collapsible UI block.

• 📎 Rich Document Handling:

  • Import: Seamlessly read from attached .pdf, .txt, and code files (.py, .json, .csv, etc.), as well as images.

  • Export: Save AI responses as beautifully formatted .pdf or raw .txt files with one click.

• 💾 Persistent Chat History: Automatically saves your chat sessions locally as .json files, accessible via the sidebar.

• 📦 Professional Setup Wizard: Includes PowerShell scripts to automatically compile the Python code into an .exe and wrap it in a standard Windows Setup Installer using Inno Setup.

🛠️ Prerequisites

1. LM Studio: Must be installed and running on your network.

2. Python 3.10+: Make sure Python is added to your system PATH.

3. Inno Setup 6 (Optional): Only required if you want to build the distributable setup installer.

🚀 Getting Started

1. Setup LM Studio

1. Open LM Studio and load your preferred model (e.g., DeepSeek R1, Qwen Coder, or Llama 3).

2. Go to the Local Server tab.

3. Ensure the server is running on http://localhost:1234/v1.

4. Click Start Server.

2. Setup the Chatbot

Clone this repository or download the source code, then install the dependencies in PowerShell: pip install -r requirements.txt

3. Run the App

Launch the app using the provided PowerShell script: powershell -ExecutionPolicy Bypass -File .\chatbot.ps1

📦 Compiling to an Installer

Want to share the app across different machines without dealing with Python dependencies? You can compile the entire project into a standard Windows Setup Installer.

1. Ensure Inno Setup 6 is installed on your machine.

2. Run the build script in PowerShell: powershell -ExecutionPolicy Bypass -File .\build.ps1

3. The script will gather dependencies, compile the .exe, and generate the installer.

4. Once finished, navigate to the newly created dist_installer folder and run LocalChatbot_Setup.exe.

📂 Project Structure

• app.py - The main CustomTkinter application logic and UI.

• requirements.txt - Python dependencies (openai, customtkinter, psutil, pynvml, etc.).

• chatbot.ps1 - PowerShell quick-launch script that verifies the LM Studio server.

• build.ps1 - PowerShell automation script to compile via PyInstaller & Inno Setup.

• chatbot.spec - Configuration file for PyInstaller executable bundling.

• setup.iss - Inno Setup configuration file for generating the installation wizard.

ℹ️ Project Info

• Version: 1.0

• Author: Arun Verma (arun.verma@iitdh.ac.in)

• Repository: github.com/arunniperk/LMStudio_chatbot

🤝 Dependencies & Acknowledgements

• UI built with CustomTkinter.

• LLM API connections powered by the openai Python client.

• Telemetry driven by psutil and pynvml.

• PDF operations handled by pypdf and fpdf2.