OllaPy
Private, Self-Hosted Interface for Ollama

OllaPy is a sleek, self-contained web interface for interacting with local language models through Ollama. It provides a private and secure environment for your conversations, running entirely on your local machine. With a lightweight Python/Flask backend and a vanilla JavaScript frontend, OllaPy is designed for simplicity, privacy, and hackability.

📖 Table of Contents

• Features
• Architecture
• Getting Started
  • Prerequisites
  • Installation & Usage
• Configuration
• API Endpoints
• Contributing

---

✨ Features

• 💻 Desktop Application: OllaPy can now be run as a standalone desktop application using Electron, providing a native-like experience.
• 🔒 Absolute Privacy: All interactions happen on your local machine. No data is ever sent to third-party servers.
• 💾 Chat History: Conversations are automatically saved as JSON files in a logs directory, allowing for easy access, backup, and management.
• 🤖 Model Selection: Seamlessly switch between different Ollama models using a dropdown menu in the user interface.
• ✍️ Markdown Rendering: Enjoy beautifully formatted AI responses, including lists, tables, and code blocks.
• 💨 Real-Time Streaming: Experience interactive conversations with the AI's responses streamed in real-time.
• 📊 Token Counter: Monitor the context size of your conversations with a progress bar and token counter.
• ⏱️ Performance Metrics: Track the generation time for each AI response.
• 📈 System Monitoring: Real-time display of CPU and RAM usage within the Electron application.
• 🚀 Integrated Backend Startup: The Electron application now automatically manages the startup and shutdown of the Flask backend server.
• 🛑 Cancel Responses: Interrupt the AI's response generation at any time.
• 🛡️ Built-in Security: Client-side HTML sanitization using DOMPurify to prevent XSS attacks.

🖼️ Screenshots

🛠️ Architecture

OllaPy's architecture is composed of two main components:

1. server.py (Backend): A lightweight Flask server that serves the main index.html file and provides a REST API for managing chat logs (CRUD operations).
2. index.html & JavaScript modules (Frontend): A vanilla JavaScript application that communicates directly with the Ollama server for AI interactions and with the Flask server for chat history management.

graph TD
    A[User 👨‍💻] -- Interacts with --> B[Browser: index.html]
    B -- "API Requests (Save/Load)" --> C[Backend Flask: server.py]
    C -- "Reads/Writes" --> D["Log Files<br/>(logs/*.json) 📝"]
    B -- "LLM Requests (Prompt)" --> E[Ollama Server 🧠]
    E -- "Streaming Response" --> B

Loading

🚀 Getting Started

Prerequisites

• Python 3 and pip.
• Ollama installed and running.
• At least one Ollama model downloaded (e.g., ollama pull gemma3).

Installation & Usage

1. Clone the repository:

   git clone https://github.com/your-username/ollapy.git
   cd ollapy

2. Run the start script:

   The start.sh script automates the setup process, including installing dependencies and launching the necessary servers.

   ./start.sh

   This will:

   • Install the required Python packages from requirements.txt.
   • Start the Ollama server in the background.
   • Start the Flask backend server.
   • Open the OllaPy web interface in your default browser.

3. Start chatting!

   You can now interact with your local language models through the OllaPy interface.

Desktop Application (Electron)

OllaPy can also be run as a desktop application using Electron, providing a native-like experience.

1. Install Node.js dependencies:

   npm install

2. Start the Electron application:

   npm start

   This will launch the OllaPy desktop application. Ensure your Ollama server is running in the background.

   Note: The Electron application currently expects the Flask backend to be running separately. You can start it using ./start.sh (which also starts Ollama) or by manually running python server.py in a separate terminal.

⚙️ Configuration

To customize the default settings, such as the default Ollama model or the Ollama server URL, you can modify the js/config.js file:

// js/config.js
export const DEFAULT_OLLAMA_MODEL = "gemma3"; // Set your preferred default model
export const OLLAMA_BASE_URL = "http://localhost:11434"; // Modify if your Ollama server runs on a different URL

🤓 API Endpoints

The Flask backend exposes a simple REST API for managing chat logs:

• GET /api/chats: Retrieves a list of all saved chats.
• GET /api/chats/<chat_id>: Retrieves the content of a specific chat.
• POST /api/chats: Saves or updates a chat.
• DELETE /api/chats/<chat_id>: Deletes a specific chat.

🤝 Contributing

Contributions are welcome! If you have any ideas for improvements or new features, feel free to fork the repository, make your changes, and submit a pull request.

Some ideas for contributions include:

• Implementing a more advanced chat history with folders and search functionality.
• Adding support for different themes (e.g., light, dark, cyberpunk).
• Adding the ability to export chats as Markdown or PDF files.