Skip to content

Count page views and project clicks with PostHog (EU servers)? It loads only if you choose Allow. Privacy

Speech to Text live preview
All projects

Speech to Text

Transcribe a short voice take in the browser without paying for a transcription service.

Transcribing meeting notes required a paid service, even though browser speech recognition APIs already existed with no friendly interface built around them.

Tech Domain · Book Dev|Published|Evidence: Live|Next.js, TypeScript, Web Speech API

A public deployment is available to try.

Case Study

Before

Transcribing short voice notes required paid services or manual typing. Browser speech recognition existed but had no simple interface around it.

After

Speech to Text on the App shell: speak and watch the transcript appear, then copy it — with no API key or server of its own.

How it's solved

Built a recorder on the Web Speech API (SpeechRecognition) with continuous recognition and interim results: start, stop, record again, read the live transcript with a word count, copy it, and clear it. It shows a clear message when the browser does not support it or the microphone errors.

Trade-offs

Recognition runs in the browser's speech service (Chrome sends audio to its provider), so support and accuracy vary; no language picker, export or saved history — the transcript is gone when you leave the page.

Related Articles

Further reading on related topics from the blog

Related Projects

More in Tech Domain · Book Dev