Sovereign GPT

This VS Code extension allows you to chat with DeepSeek-R1 and other models locally, leveraging Ollama for AI-powered responses.

Features

• Local AI Chat: Interact with DeepSeek-R1 directly in VS Code.
• Multiple Models: Choose between DeepSeek-R1, Qwen 2.5, and LLaMA 3.2.
• Real-time Streaming: Responses are streamed dynamically.
• Integrated UI: Clean and responsive chat interface.

Available Models

Model Name	Description
deepseek-r1:1.5b	DeepSeek R1 (1.5B params)
qwen2.5:1.5b	Qwen 2.5 (1.5B params)
llama3.2:1b	LLaMA 3.2 (1B params)

Installation

Prerequisites

• Install Ollama to run DeepSeek-R1 and other models locally.

Manual Installation (via .vsix)

1. Build the extension by running:

   npm install
   npm run compile

2. Package the extension:

    vsce package

3. Install the .vsix file in VS Code:

    * Open Extensions (Cmd+Shift+X).
    * Click ... (top-right) → Install from VSIX...
    * Select the generated .vsix file.

Usage Instructions

1. Open **Command Palette** (`Cmd+Shift+P` / `Ctrl+Shift+P`).
2. Run `"Start DeepSeek Chat"`.
3. Type your query and receive AI-generated responses.

Development

Build and Run Locally

npm install
npm run watch

To test in VS Code:

code --extensionDevelopmentPath=.

Linting

npm run lint

Running Tests

npm test

License

This project is licensed under the **MIT License**.