This repository contains the code for an audio transcription service using Flask and the speech_recognition library. The service accepts audio files, transcribes the speech to text, and returns the transcription as a JSON response.
- ESP32 S3 (Successfully Tested)
- ESP32 DOIT Devkit V1 (Under Test)
- Faster and LightWeight
- No need Of any Accounts,Credit Card or API key.
- Upload audio files for transcription.
- Uses
speech_recognitionlibrary for speech-to-text conversion. - Simple Flask web server to handle file uploads and return transcriptions.
- Python 3.7 or later installed.
pippackage installer.- Internet connection for installing dependencies.
- ESP Version 2.0.14
- Arduino IDE 2.3.2
- I2S MIC (INMP-441)
- download and install python if not installed from here
- open CMD on windows [Linux user open terminal & Run Equivalent Commands]
-
pip3 install flask
-
pip3 install SpeechRecognition pydub
-
Step1: Download stt_api.py and see the directory of the downloaded file
-
Step2: Open CMD Then Goto that folder and Run command
python stt_api.py- Step3: Download 'SpeechToText_ESP32' and open in Arduino IDE. Set Wifi Credentials,IP address of the server and STT timer.
- I'm Using INMP441 & ESP32 S3 wroom-1 N16R8. [Note: Use your own board and I2S Pins]