codestinymake it happen

Codestiny tools / in development

Mindump.

Say it now. Come back to it later.

Capture thoughts with your voice or in writing. Keep recordings, transcripts and your own edits together — from the first sentence to the finished document.

Check availability
MINDUMP

Thursday · an idea for later

00:12

I will capture this thought first. I can give it structure when I have the space.

Workflow illustration, not an interactive player.

Capture a thought

Record with transcription or keep audio alone. Write a note or import an audio file you already have.

Keep the connections

Browse recordings and their related texts. Return to the source, choose a model for another transcription and retain previous results.

Edit and export

Edit text, assign speakers and adjust timestamp display. Export to DOCX, PDF, Markdown, TXT, JSON, SRT or VTT.

Your audio

Local models. An explicit choice.

The desktop application is being developed primarily for local files and models. Parakeet is the default engine; Whisper requires selection. This is not a promise of zero network traffic: model downloads are a separate operation, and browser speech capabilities depend on the environment.

Can I download Mindump yet?

There is no public, verified installer yet. The desktop application is in active development and testing. This page does not present a development build as a stable release.

What about macOS speech recognition?

Local Apple Speech integration is being tested. Support depends on the system and language; Polish was not available locally through this interface on the tested Mac. Parakeet remains the primary path for Polish.

Will there be web, mobile and Gabinet versions?

Web and mobile implementations exist, but device testing and feature parity are not complete. Gabinet is a planned variant of the shared application focused on transcript work, without cloud synchronization of recordings or notes.

Status / development version

Reliability first. Release next.

Before public downloads, we are completing tests for long recordings, recovery, audio processing and a signed macOS installer. A release date has not been set.

What comes next

Next steps

  1. 01

    Fast transcription startup and verified behavior for long recordings, interrupted sessions and low disk space.

  2. 02

    Clear enhanced-audio variants, improved editing and transcription quality measurements using the same samples.

  3. 03

    A verified macOS release, followed by web/mobile and other operating-system testing.