Upload an interview, call, lesson, or voice note.
Audio intelligence for focused teams
Keep every spoken idea searchable.
Krasvi is a clean starting point for a transcription SaaS: upload audio, process it with your provider, and persist the transcript in a database without tying the product to one vendor.
- First transcript
- < 2 min
- Supported flow
- Audio to notes
- Storage model
- Postgres-ready
Board meeting audio
38:24 min · English · queued
speaker diarizationprovider: mock
Transcript
“We need a faster way to turn long conversations into decisions, owners, and reusable context.”
Persistence
- status: completed
- language: en
- storage: Postgres
Product path
A small architecture that can grow.
Send it to your chosen transcription provider through one API layer.
Save the transcript, metadata, and processing state in your database.
Search, reuse, and export the result from one quiet workspace.
Stack
Next.js for the product, adapters for the heavy work.
The repo ships with API boundaries for health checks and transcription jobs. Database and provider credentials are optional until you connect the real services.
Web
Next.js App Router, TypeScript, Tailwind CSS, standalone Cloud Run build.
Transcription
Provider interface with a mock implementation and an HTTP provider slot.
Persistence
Postgres schema and repository wired through DATABASE_URL when ready.