Productivity / Speech

Speech Transcription

Some ideas arrive more honestly in voice than in text. I built this to catch those moments before they disappear, then return to them later with less friction.

Back to workOpen live product
Role
Owner / Product Engineer across product scope, mobile capture entry, recording state, local data contract, and export handoff.
Time
Ongoing · 2026-04-04
Platforms
Web

A quieter way to hold spoken thoughts

01 / Background

Why this exists

Transcription tools win on instant usability and reliable retention. This product keeps controls minimal and state feedback explicit.
Speech Transcription current product surface
Current product surface

The current public Speech Transcription surface.

01

02 / Scenario

Use scenarios

  • Capture spoken ideas
  • Start quickly from a phone home screen
  • Hand exports into a local knowledge workflow
01Quick capture entry
02Audio + transcript retention
03Bundle export into a local workflow
Project record

The project’s current priorities and delivered decisions.

Speech Transcription screenshot — 1.png
Screen 01

Interface screenshot 01.

02

03 / Delivery

What I shipped

  • Home and /capture recording entries
  • Real-time transcription with local audio and text retention
  • History, detail, copy, and single-file Bundle export
01Record
02Transcribe
03Save
04Review
Product flow

The core path from input to outcome.

Speech Transcription screenshot — 2.png
Screen 02

Interface screenshot 02.

03

04 / Design

Design decisions

  • State clarity matters more than feature count.
  • History is treated as the core long-term asset.
04

05 / Tech

Implementation

  • Next.js 16 / React 19 / TypeScript / Zustand
  • Web Speech API transcription; localStorage metadata and IndexedDB audio blobs
  • PWA manifest and a Bundle handoff to the local OpenClaw importer
05