App Page
Audio in, structured notes out.
Upload audio. Get transcript, speaker identification, key takeaways, action items.
Why I built it
I wanted voice notes and recordings to end as useful structure, not walls of transcript text.
Problem
I record a lot of voice notes. Ideas while driving, meeting notes, thoughts I don't want to lose. But raw transcripts are just noise. Hours of speech become hundreds of lines that don't help. What was actually said? Who said it? What matters? Yes, Google and Microsoft have transcription. But I wanted to experience building a simple product myself — and I wanted the output to go beyond a wall of text into something actually structured and useful.
Value
Upload a file. Get back: clean transcript with speaker identification, key points extracted, summaries of what was discussed, action items if they exist. It's not transcription for the sake of transcription. It's transcription for understanding. A meeting becomes a useful summary. A brainstorm becomes a list of ideas. A recorded conversation becomes quotes and themes I can actually reference.
Build
Build 01
Audio upload into structured output, not a raw transcript dump.
Build 02
Speaker identification and diarization.
Build 03
Key takeaways and action item extraction, with everything saved to the knowledge base.
Preview
Ecosystem
Transcriber is another input to the knowledge base. Combined with PO (highlights), Indy Jones (web pages), and GitHub Reader (code understanding), it's a complete system for capturing everything and making it searchable through Zara.
