This project transcribes Japanese Romaji to IPA, and English to IPA (with support from @mphilli's IPA Generator), and then calculates the discrepancy using edit distance, and finally labels the indices for each phoneme.
For example, if you give such an input:
basuketto basket
The output would be like this:
basuketto, bæskət, b1a2s3u4k5e6t7t8o9, b1æ2s3k5ə6t8, 2.a/æ 4.u/ø 6.e/ə 7.t/ø 9.o/ø