Minime App

• Drag and drop input. Files can be added by dragging them from the system file manager onto the application window or by using the native file selection dialog. The interface accepts multiple images at once and queues them for processing without blocking interaction with other controls.
• Batch image conversion. Several images are processed sequentially in one session. The application keeps the interface responsive while encoding runs, and each file is handled independently so that an issue with one source does not stop the remaining queue.
• Language switching. The interface is available in English and French. The default language is English, and users can switch to French from the settings. All menus, labels, and status messages are translated.
• Theme selection. The application can follow the operating system theme automatically. Users can also force a light or dark appearance. This setting changes the color scheme of the entire window, including panels, buttons, and text.
• Original and result comparison. Before saving, a side-by-side view shows the source image and the generated output. Users can inspect the visual result at full resolution and confirm that no pixel-level changes occurred during encoding.
• File detail inspection. After processing, the interface displays the format, width, height, input file size, output file size, and the amount of saved space. This data is shown per file and helps assess the result of compression or conversion.
• Automatic smallest representation. Auto mode evaluates all supported lossless output formats for a given source and selects the one that produces the smallest file while preserving exact decoded pixels. No quality reduction is applied at any stage of this search.
• PNG effort control. The encoder provides fast, balanced, and maximum effort levels for PNG output. Higher effort increases the time spent searching for a smaller compressed stream but does not alter the decoded image content.
• Custom output folder. Converted images can be saved next to their originals or written to a separate directory chosen by the user. The path is stored for the current session and can be changed at any time.
• Larger file rejection. An option prevents writing an output file when its size exceeds that of the source image. This avoids replacing a smaller original with a larger converted version.
• ICC profile preservation. When the output format supports embedded color profiles, the ICC data from the source image is copied into the generated file. PNG, WebP, and TIFF outputs receive this treatment. Profiles are not stripped during conversion.
• Pixel verification before saving. Each output image is decoded and compared to the decoded source image. Dimensions and every 16-bit RGBA value must match exactly. If verification fails, the file is not written to disk.
• Non-overwriting file naming. Generated files use names such as photo.minime.webp or photo.minime-2.webp to avoid replacing existing files. The application checks for collisions and appends a numeric suffix when necessary.
• Animated image rejection. Animated GIF, animated WebP, and APNG inputs are not silently flattened. The application detects animation and refuses to process such files, preventing unintended loss of frames.
• Update checks. The application can check for new releases manually or automatically once per day at startup. Automatic checks are enabled only after user consent. Installing an update always requires a separate user action.

The project began as a private experiment focused on strict lossless image conversion. The developer, known under the GitHub username scorpion7slayer, built the application in Rust. The GPUI framework was used for the desktop interface. The first public release appeared after the core encoder and verification routines had stabilized. Work continues on format support, platform packaging, and update delivery.