Up till now Lojbanists had the following three main tools to search for rhymes for Lojban words:
3. own human memory
Here is a new tool that is a part of the new version of la sutysisku.
To find rhymes add @ at the end or in the beginning of the word.
It employs an algorithm written from scratch and exceeds the first two tools in several ways:
a. words with more precise rhymes are shown higher in the list.
b. phonemes closer to each other in their articulation are taken into account. Thus for words ending in -e words ending in -i have a higher priority that those ending in -u. Also {l} is similar to {r} so {saclu} rhymes with {bacru} because {l} is close to {r}. With lower priority {saclu} rhymes with {cacra} because the last letter is different.
Thus, rhymes to {saclu} are richer than those provided by earlier rhyme lists.
We only had {kacma} and {racli} as suggestions. Now we have many more.
Also note that outputting word combinations (that is fragments of text) as results is not currently implemented as this first requires writing better parsers of Lojban text than what is currently provided by PEG parsers.
Bonus:
Interesting that when working on the algorithm for this app (to be used for art purposes) another art project was born. It's an orthography more strictly enabling audio-visual isomorphism:
It's both an art project and a part of the algorithm for searching rhymes.