We've added 'bedrock' as a first-class AI provider, allowing users to run evaluations using models hosted on AWS Bedrock. By leveraging the standardized AWS credential chain (SSO, IAM roles, and env vars), this integration eliminates the need for individual API keys, making it much easier to use in locked-down corporate environments. It supports both native model IDs and inference profile ARNs with robust error handling for common AWS authentication issues.

This release adds robust support for catalog-based local testing, allowing users to download external testcase bundles via the new grindx --fetch-testcases command. We've also streamlined the testing workflow with new keyboard shortcuts: Ctrl+R to run tests and Ctrl+Shift+R to reset the timer. For custom problems, please note the transition to a per-directory structure under ~/.grindx/problems/, with previous custom.json support now deprecated. Enjoy a much smoother development loop with these workflow enhancements! 🚀

This update introduces a new catalog-based system for managing problem metadata and testcases, decoupling them from the main codebase. Developers can now use grindx --fetch-testcases to download problem assets, enabling a more modular and scalable approach to local test execution. A new folder-based structure for custom problems has also been adopted to improve organization.

This update introduces several workflow improvements, headlined by AI auto-review that validates solutions upon completion. We've also added a search feature (trigger with /) to easily find problems, an activity calendar on the welcome screen to track your progress, and improved CLI support for listing solved problems. These additions, combined with various UI and bug fixes, make tracking your grinding sessions significantly smoother.

This release adds the requested ability to define custom problem sheets and individual problems by dropping files into your ~/.grindx directory. We've also addressed several reliability issues, including switching to atomic file writes for progress tracking to prevent corruption, fixing language detection across all supported runtimes, and improving editor behavior. You can now use the new --list-problems flag to better manage your workflow.

Updated the terminal application to save your chosen programming language in the progress configuration file. Previously, the language reset to the default on every session, requiring manual adjustment. The solve screen now automatically initializes with your saved preference, ensuring a smoother start to your practice drills.

grindx now supports C++, Java, and JavaScript for DSA practice, alongside existing Python and Go support. Users can now easily switch between these languages using Ctrl+L within the solve screen, making it even more versatile for multi-language interviews and algorithm prep.

Grindx now supports a wider range of programming languages for DSA practice. We've updated the core editor and problem structures to allow seamless switching between Python, Go, C++, Java, and JS using the Ctrl+L shortcut, providing more flexibility for your preferred environment.

Added a new GitHub Actions workflow to automate package releases to PyPI. The pipeline is configured to trigger automatically upon pushing a tagged release, ensuring consistent deployment of build artifacts. This streamlines the release process by removing manual upload steps while utilizing OIDC for secure authentication.