Scriberr

Ever wished you could just record something and get an accurate transcript with speaker identification? That's exactly what Scriberr does.

What is this?

Scriberr is a web app that takes your audio files and turns them into detailed transcripts. It uses WhisperX under the hood, so you get:

• Accurate transcription - WhisperX is really good at this
• Speaker diarization - It figures out who said what
• Multiple formats - Get your transcript as text, SRT, VTT, or JSON
• Chat with your audio - Ask questions about what was said using AI
• Take notes - Annotate important parts as you listen

Quick start

The easiest way to get started:

brew tap rishikanthc/scriberr
brew install scriberr

Then just run scriberr and open http://localhost:8080 in your browser.

What you can do

• Upload audio files or record directly in the app
• Get transcripts with timestamps and speaker labels
• Download transcripts in whatever format you need
• Chat with AI about your recordings
• Create summaries of long audio
• Manage everything through a clean web interface

Requirements

• Python 3.11+ (for the transcription engine)
• A few GB of disk space for the AI models

That's it. Everything else is handled for you.

Built with

Go backend, React frontend, WhisperX for transcription, and a lot of coffee.