CaptainCoder

🚧 Work In Progress 🚧

Note: This project is currently a work in progress and is not yet finalized.

Description

CaptainCoder is a privacy-focused, locally-run code completion extension for VSCode that supports a wide range of popular large language models. With open-source flexibility and the integration of Retrieval-Augmented Generation (RAG), it provides enhanced context-aware suggestions and custom embedding models, ensuring advanced, tailored autocompletion while maintaining full control over your code and data privacy.

Features

• AI-Powered Code Completion: Utilizes customizable models (Ollama, OpenAI, Mistral, Hugging Face, Anthropic, etc.) for intelligent code suggestions.
• Privacy and Control: Operates locally to ensure complete privacy and control over your code, made possible by local support of Ollama and Retrieval-Augmented Generation (RAG) context.
• Open-Source and Extensible: Fully open-source, allowing for community contributions and extensions.
• Extensible Model Provider: Offers support for a variety of popular large language models, both paid and open-source, allowing you to choose based on your privacy and budget preferences.
• Context-Aware Suggestions: Provides enhanced autocompletion with a deeper understanding of the entire codebase by supplying the model with complete code context via RAG.
• Custom Embedding Models: Allows users to create and use custom embedding models for specialized use cases by feeding various types of documents (e.g., PDF, web pages, Microsoft documents, Confluence pages, text files, JSON, CSV, etc.) to the models through RAG.
• Flexible Model Switching: Enables easy switching between different models while retaining customized enhancements and specialized functionalities, ensuring adaptability to various needs.

Installation

1. Install Ollama Locally:

   • Visit https://ollama.com/ for detailed installation instructions.

2. Install CaptainCoder VSCode Extension:

   • Open VSCode.
   • Go to the Extensions view by clicking on the Extensions icon in the Activity Bar on the side of the window or by pressing Ctrl+Shift+X.
   • Search for CaptainCoder.
   • Click Install to add the extension to your VSCode setup.

Usage (Currently Only Basic)

1. List Available Models:

   • Use the command CaptainCoder: List Models to fetch the list of available models.
   • In the Command Palette (Ctrl+Shift+P), type CaptainCoder: List Models and press Enter.
   • A notification will display the fetched data, and the list will be logged in the console.

2. Install New Models:

   • Use the command CaptainCoder: Pull Model to install a new model.
   • In the Command Palette (Ctrl+Shift+P), type CaptainCoder: Pull Model and press Enter.
   • Enter the name of the model you want to download when prompted.
   • A progress notification will show the download status, and a message will appear upon successful download.

3. Pick Project Files:

   • Use the command CaptainCoder: Pick Code Content to select code files from your project.
   • In the Command Palette (Ctrl+Shift+P), type CaptainCoder: Pick Code Content and press Enter.
   • The content will be saved to a file named code_content.txt in the root of your workspace.

4. Load Vector Store:

   • Use the command CaptainCoder: Load Vector Store to load the vector store with project content.
   • In the Command Palette (Ctrl+Shift+P), type CaptainCoder: Load Vector Store and press Enter.
   • A notification will confirm the successful loading of the vector store.

5. Chat With The Captain:

   • Use the command CaptainCoder: Chat With The Captain to interact with the chat interface.
   • In the Command Palette (Ctrl+Shift+P), type CaptainCoder: Chat With The Captain and press Enter.
   • A chat panel will open. Type your question in the input box and click Send.
   • The response will be displayed in the chat panel.

Issues

If you encounter any issues, please report them on the GitHub issues page.

Contributing

Contributions are welcome! Please fork the repository and submit a pull request.

License

This project is licensed under the MIT License.