Thank you to everyone who has helped improve ChessVision. This file recognizes the people who have contributed code, fixes, and improvements to the project.
Khatai Huseynzada — @BilgeGates
Project author and maintainer.
Last contribution: May 2026
A hidden canvas render pipeline that ran continuously in the background — separate from the export process — was identified and removed. The board now renders only where needed, reducing idle CPU usage.
A repetitive draw cycle on the main page that increased render load during navigation and settings changes was simplified, resulting in smoother in-app transitions.
A scaling miscalculation in the 24× and 32× export modes caused the output dimensions to deviate from the specified board size in centimetres. The formula was corrected so that, for example, a 4 cm board at 32× quality now produces output based on that exact measurement.
The logic for computing export dimensions was consolidated into a single source. Previously, the dimension displayed in the UI could differ from the dimension used to generate the canvas. This inconsistency has been resolved.
The drag preview for pieces was misaligned from the cursor during drag operations. The preview now tracks the cursor accurately throughout the drag interaction.
Asset caching improved — #53
Repeated loading of identical piece images across pages was reduced. This lightens the load during page transitions and general use.
The progress bar shown during export now reflects actual processing steps rather than advancing at an arbitrary rate.
PNG, JPEG, and SVG export flows were refactored for consistency and reliability.
In multi-position export, each position is now downloaded sequentially with an incrementing numeric filename.
For 24× and 32× quality options, the interface now displays the expected output dimensions so users can anticipate file size before exporting.
A warning is shown when an export configuration may place significant demand on the device, allowing users to make an informed choice before proceeding.
Validation errors now indicate the specific part of the FEN string that failed, rather than returning a generic invalid message.
Users can now export their saved data, restore it from a backup, or reset it entirely from within the settings panel.
Basic automated tests were introduced for core FEN validation logic to help catch regressions.
Internal documentation was updated to reflect the optimisations made and to note considerations for large export operations.
Last contribution: May 2026
An issue causing saved user data — including FEN history, favourites, and settings — to be unexpectedly cleared from localStorage was identified and fixed.
A maximum length constraint was added to the FEN input field to prevent excessively long strings from being submitted, protecting against potential parsing issues and UI instability.
To contribute to ChessVision, please read CONTRIBUTING.md and open a pull request.