Starting an audio tool from the share sheet

When the useful part of an audio clip is a sentence or an idea, reading can be easier than replaying it. I built an early iPhone prototype around that task: share a file, process its audio and keep a transcript you can return to.

The starting point was the share sheet. Rather than asking someone to find the app first, the design accepts an audio or video file from another place on the phone.

Keep the flow close to the source

A share extension hands the file to the app through a shared inbox. The main app imports it, extracts the audio when needed and prepares it for the speech engine.

The interface includes a transcript library and a detail view. Transcript segments carry timestamps so the text can retain a connection to the recording.

Build around the device

The prototype uses native iPhone frameworks for the interface, local storage and speech processing. Text exports support plain notes, Markdown and subtitle files.

The choice was to start with the platform’s speech engine and keep a different model as a later option only if real recordings showed a useful improvement. That keeps the initial product smaller and makes the quality question something to test.

Separate a working interface from a tested product

The project produced an app scaffold, share extension, library, export path and simulator screens. The exporter had focused checks, but the core transcription path still needed a real-device run. Speech assets were unavailable in the simulator used for the first pass.

That limitation matters: a successful build can establish that the pieces fit together without establishing how well the complete experience works for someone holding a phone.

The next useful test

Try the full import-to-export flow on a real device with representative recordings. Review transcripts against human-checked references and exercise the share extension from the apps people actually use.

This is an early prototype, not an App Store release or a measured transcription-quality claim. The illustration is a conceptual view of audio becoming readable notes.