Overview
TransmuSynth takes the concepts of two other utilities I wrote, image2sound and prompt2pixel, and combines them into one. Users can either input text or generate random text, and TransmuSynth will use various algorithms to convert the text into an image. That image is then converted into MIDI, using different parameters, such as key and tempo, which the user can specify. Alternatively, users can provide their own images to convert to MIDI.
The name TransmuSynth is a portmanteau of transmutation and synthesis. The application is sort of like "media alchemy" as text becomes a cryptographic hashsum, which becomes an image, and the image becomes MIDI (originally WAV files in the CLI version).
Check out the Usage Guide for an explanation of the various options available for each tool.
Version Information
The current version of TransmuSynth is 20260507. As it is still being developed, bugs of all sorts are likely. Feel free to report them on Codeberg.
It was developed from image2sound version 0.12 and the development version of prompt2pixel. (No releases of the latter utility have been tagged at this time.)
License
TransmuSynth is © 2026 by John Beers and is made available under a GNU AGPL 3.0+ license