Helen is a good english teacher and can help you improve your speaking habits. Helen provide some features that can help during your speaking time.

Helen is beta project. I wish I make the project to be visualized using typescript and next.js. That is my goal. So, I'm going to write the code using Typescript and Next.js and langchainJs.

You can record your voice when you are talking with someone. The voice file will be located resources folder. The voice file will be separated by 3 minutes.

[folder structure]

The structure of resource folder follow as below structure

- resources
    - date (yy-mm-dd)
        - audio (voice files)
            - original (your voice file)
                - 23-05-13.wav
            - voice files which was sparated by minutes. (3 minutes)
                - 23-05-13_0.wav
                - 23-05-13_1.wav
        - transcript (transcript file was maded from your voice file)
        - review (To be fixied by chat-gpt)

We'll fix the feature that can record system sound. Because, I think we need to separate voice file to each person's sentence and separate the voice to make transcript. I think we'll use the hugging face to separate voice.

If you recorded your voice perfectly, you can make transcript file from your voice file. If you execute the main.py in transcript folder, then It make the transcript file from your voice file.

[Transcript file]

NLP feature can separate the sentence to each sentence. The sentence will be put in GPT per each sentence.

Correction feature can fix the sentence that you spoke when you were talking someone. The correction format must be bellow format. You can study from the format.

Format

- Original sentence (written by the student):
- Grammatically correct sentence:
- Explanation (in bullet point form):
- Example conversation (briefly):

Prerequisites

• You must have python 3 or later installed.
• You must install poetry
• Make .env file on the root directory of project and write your api-key

[env file]

OPENAI_API_KEY=

1. clone the repository

git clone https://github.com/tmdgusya/helen.git

2. Install dependencies

poetry install

3. Record your voice

poetry run python src/audio/audio.py

4. Resize your voice file to small

poetry run python src/audio/resizer.py

5. Make Transcript from voice file

poetry run python src/transcript/whisper.py

6. Fix your sentence

poetry run python src/main.py

7. Check the file that written by Open-AI

• I wish I make the program to visualize using typescript and next.js
• I wish I separate the transcript to each person's sentence