VoiceScribe AI — original audio-to-ChatGPT demo
Verified: 2026-11-07

This package contains an original 17.491-second synthetic recording and real
product exports. The voices were generated with macOS Samantha and Daniel.
This is not customer audio, a customer testimonial or an accuracy benchmark.

voicescribe-podcast-demo.wav / .mp3: identical spoken content in two formats.
reference-script.txt: the original authored text used to synthesize the audio.
original-transcript.txt / original-subtitles.srt / original-subtitles.vtt:
  unchanged VoiceScribe exports downloaded before manual speaker correction.

Known original-output issue: the standalone word "I" at approximately 00:07
was assigned to speaker_0 rather than the beginning of speaker_1's sentence.
The subsequent editor screenshot shows a real, saved manual correction.
The exports here preserve the earlier output and have not been silently edited.

To reproduce, upload one audio file to your own VoiceScribe account, wait for
completion, review and save edits, then connect that account to ChatGPT.
AI summaries and transcripts beyond the 30-second free preview require Pro.
A 7-day Pro trial requires payment details and is followed by a paid subscription. ChatGPT app access depends on
your account and workspace settings. Review the permissions when connecting.

VoiceScribe: https://voicescribeai.net
Setup: https://voicescribeai.net/integrations/chatgpt
Support: https://voicescribeai.net/contact
