I made an algorithm to vectorize the phonetics of a word. The idea is to be able to find words that share sounds anywhere, not just at the beginning (alliteration) or the end (rhyme) . The human ear can detect all kinds of sound repetition, so why can’t my computer?

Below are examples of different search tasks using the phonetic vectors I made.

Search by word:
skull :: Scully, cull, scuttle, sculpted, scholar, scum

Constrain by meter:
skull <‘1’> :: cull, sculpt, scum, hull, skill, scrum, school, scar
skull <’01’> :: succumb, annul, succumbed, occult, cabal, discuss, recall
skull <’10’> :: Scully, scuttle, sculpted, scholar, scuffle, scurry, scarlet, Scotland
skull <‘100’> :: sculptural, scuttling, colorful
skull <‘010’> :: discolored, expulsion, succumbing, discomfort

Combine with another word:
skull + blue :: slew, bluegill, hullabaloo, Sue, clue, skew, screw
skull + green :: screen, gull, scorn, gruel, skein
skull + yellow :: skeletal, yell, slow, sallow, smell
skull + red :: sled, scared, screwed, revel

Combine with word meaning:
a [color word] sounding like “skull” :: discolored, scarlet, teal, sparkly, scalloped, speckled

Search by specific sound(s):
(starts “sk”, ends “l”) :: scale, school, scowl, schedule, scuttle, squirrel, skeletal, scramble, scrap metal
(“ull”) :: annul, gully, multicolored, cull, gullible, adult, ultra, sulk, cult, bulge
(“skl”) :: briskly, disclose, exclude, skillful, scalpel, cyclical, scale, skill, escalate
(starts “sk”, ends “l”, “uh”) :: scuffle, scuttle, squirrel