AutoDev is an AI-powered coding assistant for IntelliJ IDEA, designed to enhance developer productivity through automated code generation, testing, documentation, and debugging. It targets developers seeking to streamline repetitive tasks and improve code quality by leveraging large language models.

How It Works

AutoDev employs an "Agentic drive coding workflow" using "Sketches," which are IDE-integrated tools for specific tasks like code editing, diff viewing, terminal interaction, and web preview. It supports context-aware code generation for various patterns (e.g., Controller, Service, Repository) and can automate tasks like unit test creation, bug fixing, and documentation generation. The architecture allows customization of LLM servers, prompts, and even custom AI agents.

Quick Start & Requirements

• Install via IntelliJ IDEA Plugin Marketplace.
• Requires IntelliJ IDEA.
• Additional plugins may be needed for specific "Sketches" (e.g., OpenAPI, Dependency Checker).
• Supports multiple languages including Java, Python, Go, Kotlin, JS/TS, C/C++, C#, Rust.
• Links: English Demo, Chinese Demo, VSCode Version

Highlighted Details

• "Sketch" based IDE canvas for interactive development.
• Automated code generation for common patterns (e.g., AutoCRUD, AutoSQL, AutoPage).
• AI-driven code review, refactoring, commit message generation, and CI/CD config creation.
• Customizable LLM integration and prompt engineering.

Maintenance & Community

• Project is actively developed with a VSCode counterpart.
• Mentions "Thoughtworks" as a user.
• Links: Discord/Slack, Roadmap

Licensing & Compatibility

• Core code distributed under MPL 2.0.
• Some components inspired by or based on Apache 2.0 licensed projects.
• Generally compatible with commercial use, but MPL 2.0 requires source code availability for modifications to the licensed components.

Limitations & Caveats

Some advanced features are marked as requiring additional plugin installations. The project mentions fine-tuning models on HuggingFace and OpenBayes, suggesting a focus on performance and customization but potentially requiring significant setup for optimal results.