DevoxxGenie is an IntelliJ IDEA plugin that integrates with local and cloud-based Large Language Models (LLMs) to assist developers with code review, explanation, and testing. It targets developers seeking AI-powered coding assistance directly within their IDE, offering features like RAG, Git diff integration, and support for a wide array of LLM providers.

How It Works

DevoxxGenie leverages the Langchain4J library, written entirely in Java, to provide a unified interface for various LLM providers. It supports local LLMs via Ollama, LMStudio, GPT4All, and Llama.cpp, as well as cloud services like OpenAI, Anthropic, and Gemini. A key feature is Retrieval-Augmented Generation (RAG), which uses ChromaDB and Ollama/Nomic embeddings to index project files, enabling semantic code search and contextually relevant responses.

Quick Start & Requirements

• Installation: Via IntelliJ IDEA Marketplace or by building from source (./gradlew buildPlugin).
• Prerequisites: IntelliJ IDEA minimum version 2023.3.4, Java minimum version JDK 17. Local LLMs require specific setup (e.g., Ollama installation and model download).
• Setup Time: Local LLM setup can take minutes; cloud LLM setup is typically under 2 minutes.
• Links: Video Tutorials, Blog Posts

Highlighted Details

• Supports Retrieval-Augmented Generation (RAG) with local ChromaDB for enhanced code context.
• Integrates Git diff/merge viewer for seamless acceptance of LLM-generated code changes.
• Offers multimodal support for image and file drag-and-drop with compatible LLMs.
• Includes a token cost calculator for cloud LLM usage and smart model selection based on context window and cost.

Maintenance & Community

• Active development with frequent updates and new feature introductions.
• Community engagement encouraged via pull requests.
• Links to video tutorials and blog posts are available for deeper dives.

Licensing & Compatibility

• The plugin itself is not explicitly licensed in the README. The underlying Langchain4J library is Apache 2.0 licensed.
• Compatibility with commercial or closed-source projects is likely, but the plugin's specific license should be verified.

Limitations & Caveats

• Test Driven Generation (TDG) is marked as experimental.
• The plugin's own license is not clearly stated, which could be a concern for some users.
• While it supports many LLMs, performance and integration quality may vary between providers.