I’ve been thinking about shaking up my review workflow a bit, but cannot figure out tooling to support that.
One part is fast voice recognition to replace typing.
I’ve found an old thread about this approach (LipSurf - Bunpro Voice Dictation. Review while you eat, do chores, etc), but it seems LipSurf Bunpro plugin is broken since 2023, and LipSurf itself is on life support.
Also I suppose in 2025 it should be possible to just run a Speech-to-Japanese model on a developer laptop instead of relying on laggy online services?
Second part that I’d try to integrate is “you are mistaking this with” feature.
It would show vocabulary corresponding to the user’s incorrect input, if any, to more easily remember the difference.
This could work with voice recognition + self-grading, or also could work with just typed input.
And finally these tools could be combined for a reading plugin.
The user would read some text, let’s say in ttsu, or in our favorite Haruna’s diary, and the tooling would follow along, underline parts that were read correctly, and provide “you’ve mistaken this with” popups.
And also mark any words currently due in SRS as reviewed.
Does anyone have pointers to existing tools for any parts of this puzzle?