SautiSafi

• 

  Transcription result page

• 

  Home page

• 

  Billing Page

Inspiration

Researchers in Tanzania have complained that there is poor support for transcribing audios with Swahili language compare to other languages and have been looking for solutions

What it does

This web app is able to transcribe Swahili audios with high accuracy within a relative short time

How I built it

This web app was built using NextJS for Full-stack development, MongoDB for database and google speech to text api for transcribing

Challenges I ran into

I was unable to add speaker diarization because it's not supported in google speech to text api that means i had to host another model separately that first diarize the audio then each speaker segment is transcribed separately.

Accomplishments that we're proud of

It has been launched and adopted by various people in the field

What I learned

I learned how to navigate in the Google Cloud Console, I learned how to secure Google Cloud Storage and Audio Manipulation

What's next for SautiSafi

Facilitating it's adaptation among researchers and transcripts writers

Built With

• firebaseapphosting
• google-web-speech-api
• mongodb
• nextjs