Hey all! I made an app that I’d like to share in case others find it useful.
I’ve been thinking of a way that would make context sentences more useful to me and came up with this simple exercise:
- listen to TTS narrated context sentence without knowing which particular vocabulary it uses (but only use the vocab that you learned);
- try to get the meaning, or at least to spot the familiar vocab;
- then check your guess by looking at the written and/or translated version of the same sentence.
All that, I have implemented in this simple web app: TTSensei (source code: GitHub - SiLiKhon/TTSensei ). It’s likely buggy, but should be usable.
I’d love to hear feedback, improvement suggestions, or just whether this thing is useful at all.
2 Likes
Here’s a short tutorial to get started with the app.
Setup
- Generate WaniKani API token with
all_data:read access. Paste it into the corresponding field in the settings tab of the app (“Settings” button in the top right). This will be used later on to collect all the vocabulary and kana vocabulary items for which your SRS stage is above 0.
- Generate TTS API token at apiKeyの取得 (for more info, see WEB版VOICEVOX API（高速） ). Paste into the corresponding field in the settings tab of the app. This will be used to voice the context sentences for the collected vocabulary when clicking the “►” button.
- Click “Collect Vocab & Sentences.” Optionally, you can filter by levels from which you would like the vocabulary to be taken.
This info will be stored in your browser cache, so normally you wouldn’t need to repeat these steps. You may want to collect more vocabulary and/or impose level filters (by repeating step 3 above) as you progress in WaniKani.
Usage
After the setup is done, the left panel should be filled with context sentences to pick from. Once you click on the “Get Random Sentence” button, a random sentence will be loaded, though its content will be hidden behind spoilers. You can uncover these spoilers by clicking on them. The intention is to be able to listen to the phrase (as narrated by TTS when clicking the “►” button) before you read it.
I guess the workflow would depend on your skill level a lot. For someone like myself, who is just starting to learn, I intended something like this:
- Get random sentence and listen to it (“►” button). Try to pick the familiar phrases from the sentence, even if you can’t get the entire wording.
- Check yourself by uncovering the top spoiler (vocabulary word/phrase). Listen again to cross-check that you do hear the vocabulary phrase in the narrated sentence.
- Uncover the middle spoiler (sentence written in Japanese). See if you can pick up more meaning and familiar words here. Do listen to the voice again - are you making more sense of it now?
- Finally, uncover the bottom spoiler to check the translation.
Hope this helps 
3 Likes
I made a tiny, yet hopefully useful update, by adding a link to duck.ai (anonymized chatbot) so that it makes the grammatical breakdown of the current sentence (screenshots below).
This helps me a lot at my current level, although I have no idea how trustworthy these explanations are
