Neutboom is a broader language learning MVP and as a former ESL tutor who creates their own resources (~90% do!), I used the gold-standard Forvo audio database to build it. However, I found limitations: 1) sometimes it's hard to mimic the pronunciation of someone else due to different vocal styles 2) you can't count on the database to have an audio recording for an uncommon word or language.

loxai is a solution to the above. API & MVP (to visualise user end of the process) out now.

SpeakSanDiego is an embodiment of loxai's API, but wrapped in the style of competitor BurlingtonEnglish. This is meant to bypass the familiarity issue in early user-acquisition. In a user interview, an ESL tutor suggested that users should be able to hear their pronunciation against the corrected version. This also aligns with benchmark study participant feedback that language learning methods are more engaging when they show accuracy metrics. loxai's MVP incorporates this as well.