We need your feedback! Please fill out this Feedback Form to help us improve the extension.
Features
1. AI Chatbot with RAG (Retrieval-Augmented Generation) capabilities
Document-based chatbot: AI assistant that understands the content of a specific document and can answer questions directly based on it.
General purpose chatbot: A versatile chatbot for various queries.
Beginner / developer mode: Switch between modes tailored for different user experiences.
2. Flexible LLM Engine Configuration
Seamlessly switch between local and cloud-based models to configure both the chat model and the embedding model (used for RAG feature).
Cloud
Chat models: Users can use Google Gemini model for chat and text generation tasks.
Embedding models: Users can use Hugging Face embedding models (like sentence-transformers/all-MiniLM-L6-v2) via Hugging Face Inference API.
Local
Users can use local models via Ollama or LM Studio for both chat and embedding tasks. Users can use any models supported by these local LLM frameworks, giving them flexibility and control over their AI tools.
3. Documentation Drift Analysis
Document drift detection based on Git commits: Automatically detects and highlights when your documentation becomes outdated or inconsistent with recent code changes by analyzing Git commit history.
4. Smart Documentation Tools
Generate documentation from scratch: Automatically create documentation for your projects.
Template suggestion to create documentation: Get suggestions for documentation templates.
Summarize document: Get a quick summary of your document.
Translate document: Translate your documents into different languages.
AI-generated visualizations:
Architecture visualization
Folder structure visualization
5. Editing Helper:
Grammar checking: Check and correct grammatical errors.
Markdown validators: Ensure your markdown is well-formed.
Technologies Used
TypeScript: Primary language for the extension.
VS Code API: For building the extension and integrating with the editor.
LangChain & Google Gemini: For Large Language Model (LLM) integration and RAG implementation.
Vector Databases: For storing and retrieving document embeddings efficiently.
Embedding Models: Hugging Face Transformers, Ollama, and LM Studio for semantic understanding.
esbuild: For bundling the extension.
Mermaid.js, D3.js, Vis.js: For creating visualizations.
HTML/CSS/JavaScript: For the webview-based UI components.
About This
This is the solution for CodeNection 2025 by Team JavaMee.
Track: Industry Collaboration
Problem Statement: Fix the Docs: Smarter, Faster, Maintainable Documentation for the Real World by iFAST
Team Members: Beh Shin Yeong, Chiam Huai Ren, Hoe Zhi Wan
