Whisper-API
- 5 Devlogs
- 13 Total hours
Full stack service to transcribe voice file into text
Ship #1
Changes requested
I wanted to transcribe my audio meetings into text, but everything I found was either paid or behind a subscription. So I thought, why should I pay for something I can build myself? So I did.
The real challenge was getting the model to actually run on a Raspberry Pi 5. With only 4GB of RAM it choked badly at first. I fixed it by splitting the audio into chunks and transcribing them one at a time instead of loading the whole file into memory at once.
Drop your audio, get your text back, fully self-hosted, running on a Pi.
3h 40m 36s logged
I was working on deployment and creating Vue.js dashboard
47m 2s logged
I was working on deploying to my PI 5 and testing its abilities (it did ended well). Also I added a to_email field.
1h 4m 51s logged
I finally stopped chasing perfections and shipped MVP. I mean like real MVP. I have almost no UI, but fully working transcription
6h 52m 30s logged
Just finished building MVP backend
31m 11s logged
Created base project portfolio
