Maverick

Maverick is a free, open-source Visual Studio Code extension developed by Yurts that delivers AI-powered code completion for Python. Unlike cloud-based code assistants that send code to remote servers, Maverick runs entirely on the local machine, keeping code private and enabling offline use. The extension is available through the VS Code Marketplace and is released under the BSD-3-Clause license.

Key Features

• Local AI code completion: Runs models directly on the local machine, ensuring no code is sent to external servers or knowledge bases
• Keyboard-triggered suggestions: Inline completions are triggered by pressing Ctrl+Shift+M on Windows and Linux or Cmd+Shift+M on macOS
• Cross-platform compatibility: Available for macOS (Intel and Apple Silicon), Linux (x64 and arm64), and Windows with experimental support
• Configurable settings: Users can adjust the model port, maximum tokens to generate, and number of context lines to balance inference speed and output quality
• Free and open source: Available at no cost with full source code on GitHub under the BSD-3-Clause license

How It Works

Maverick installs a local model server alongside the VS Code extension. During installation, the packaged application and code completion model are downloaded to the user machine. When the developer types Python code and triggers the completion command, the extension sends relevant code context to the local model server. The server processes the prediction entirely on the user hardware and returns an inline suggestion that appears directly in the editor. The model server runs on a configurable local port and operates without making any external network requests.

System Requirements

Maverick requires VS Code running on macOS (Intel and Apple Silicon), Linux (x64 and arm64), or Windows. A compatible GPU is not required but may improve prediction speed. Users should expect installation to take approximately ten minutes as the application and model files download. The extension requires that VS Code inline suggestions are enabled through the editor settings.

Intended Users

Maverick is designed for Python developers who want AI-assisted code completion while maintaining full control over their code privacy. It is especially useful in air-gapped environments, on projects with strict data handling requirements, or for users who simply prefer not to send their code to cloud-based AI services. The free and open-source nature makes it accessible to students and independent developers as well.

Limitations

Maverick focuses exclusively on Python code completion and does not support other programming languages. The GitHub repository was archived in October 2024, indicating the project is no longer actively maintained. Users should not expect further updates, feature additions, or bug fixes. Inference speed may be slower than cloud-based alternatives, particularly on systems without dedicated GPUs, as the model runs entirely on local hardware.