Looks up each word in the CMU pronouncing dictionary, counts every phoneme by group (stop, fricative, etc.) and acquisition age (Crowe & McLeod, 2020), and tracks where each phoneme falls in the word.

It doesn't know your client's phonology, dialect, or co-articulation patterns. It can't tell you which sounds to target — only which are present. Words outside the dictionary — usually names, invented words, or clinical jargon — are skipped and named.

Ask Claude "what sounds are in this passage?" and you'll get a confident summary with no audit trail. Here, every count is reproducible from the table — you can see the math and disagree with it.