Tambr runs a topic extraction algorithm to get groups of related words in text that define a topic. It then uses a synthesizer search engine built from human-tagged labels to select the synthezisers that best fit those topics, outputting chords and melodies that are based from the novel's evolving sentiment arcs.
Tambr is open source. Anybody can freely view and contribute to the code. If you have ideas for what timbr can become in the future, feel free to contact me or fork the repo on github.