SRS Listening App - Kikikata (Improve your Japanese Listening)

Japanese
#1

Hi everyone!

I’m a Level 60 WaniKani user. But even after mastering the Kanji and vocabulary, I found that my listening comprehension was still lagging way behind. I could read the words, but I couldn’t “hear” them in real-time.

That’s why I created Kikikata - an SRS-based listening app.

The concept is simple: You hear a sentence and type what you hear. The main difference from other apps? If you mishear a specific word, you can add that individual word to your SRS and learn it separately until you can recognize it back in the full sentence.

I posted here a while ago and received a ton of great feedback. I’m happy to share that many of those suggestions are now live:

  • Dual Study Modes: Choose between Text Input (dictation) or a Simple Mode (correct/incorrect) for faster reviews.
  • Karaoke-Style Tracking: The text highlights in real-time as the audio plays.
  • Context & Grammar: Every sentence now includes Politeness Levels, Cultural Context, and tagged Grammar Points.
  • Speed Control: Adjustable playback for audio.
  • Total UI Overhaul: A completely new look focused on a smooth, distraction-free study flow

If your listening is lagging behind your reading, I’d love for you to give it a try and share your feedback!

Check it out:kikikata.app

image
image1303×1289 196 KB

image
image912×518 28.5 KB

image
image1047×1095 101 KB

1 Like
#2

Your app shows every answer as wrong, even if I copy 1:1 what it says the answer is and enter it the next time.

#3

(post deleted by author)

#4

I’m sorry for the frustration! By default, the site is set up for Kana-only input (which is usually easier for beginners).
If you click the Session Settings (top-left corner), you can enable Kanji-Based Input. This will check your typed text against the actual Kanji instead of just the Kana.

#5

That worked, yeah.

1 Like
#6

Could you tell a little bit about your approach to sourcing content, especially for higher N levels?

I’ve looked around the landing page and docs but couldn’t find anything about either LLM, TTS or human involvement.

2 Likes
#7

+1
was about to ask the same thing. If sentences are recorded by native speakers or AI.

1 Like
#8

Currently, the content is sourced as follows:

  • Native Audio: All of the sentence-level audio (and many common words) comes from high-quality native recordings sourced from iKnow! database. I prioritized native audio over robotic Text-to-Speech (TTS) to ensure you’re hearing natural intonation.
  • TTS Supplement: For individual words where a native recording wasn’t available, I used TTS to fill the gap so you aren’t left without a reference.
  • Sentence Sources: The sentences themselves are curated from established core decks and JLPT materials across all levels, from N5 to N1.

In the future I’m looking to expand the library significantly. I plan to incorporate audio from anime (focusing on content that is accessible under fair use or specific licensing) to provide more “real-world” listening practice and I’m also exploring the latest Neural TTS models for new sentences (they have become incredibly natural recently).

I’ll make sure to update the docs with these details, thank you!

#9

Thanks, just replied above with the details