Voice Assistant Dataset Creation

I'm looking to create a dataset for voice assistant training. The dataset will be used for English language training, and will require recordings at a standard quality, clear enough for processing.

Key Responsibilities:
- Develop an online platform that records voice commands
- Show 'N' number of words [commands] for the user.
- The specific commands to be hardcoded in the script by the client.
- Record the commands for exactly for 1 second
- Repeat the same command five times. So five times can be recorded the same word from the same user [speaker].
- So 5*N commands per user to be stored online.
- Format: wav, bit rate: 256kbps, 16 kHz, Mono
- Display the 5 records, so the user can listen back, delete it if required and record new one.
- After user approval save the command (5) and repeat N times the same process.
- Save recorded wave files online in the following sub-directory "command/user name_round.wav"
For example: "light/439c84f4_1.wav"
- File name structure:
word: shown word, to be spoken by the user.
username: random (for example: 439c84f4)
round: _1: First trial, second trial: _2 etc.
- Store the summary of the wav files in the metadata.txt
For example:
...
light/439c84f4_1.wav
light/439c84f4_2.wav
light/439c84f4_3.wav
...
heater/439c84f4_4.wav
heater/439c84f4_5.wav
...
light/ff2b842e_0.wav
light/ff2b842e_1.wav
...
heater/ff2b842e_2.wav
heater/ff2b842e_3.wav
...
- Ensure the dataset is optimized for use in training
- Deliver a high-quality dataset for English voice command training
- The fully working source codes to be shared with the client

Ideal Skills:
- Proficiency in HTML, JavaScript and CSS
- Experience in audio recording and processing
- Previous work with voice command datasets or voice assistant training would be advantageous.

Use (Skills required):
MediaStream Recording API
Web Audio API

UI:
Simpe