Record each letter name once in your voice. Students do not do this — the classroom model improves from their practice judgments only. After landmark DSP updates (v0.87+), re-record this seed for each stage — old voice banks are not compatible.
0 / 26 recorded
A — say: ay
Teacher writing seed (one time)
Write each uppercase letter once in your handwriting. This trains the letter-recognition model. Students improve it through your practice judgments only.