Using Codex to Improve Voice-to-Text: Auto-Extracting Custom Vocabularies

jdjohnson · x · 2026-08-01

Shares a practical workflow to improve voice-to-text accuracy. The author suggests using Codex to analyze a user's past conversations, Slack messages, and emails to automatically extract frequently used domain-specific words and phrases into a spreadsheet. This vocabulary list can then be imported into the dictation tool Superwhisper to significantly reduce recognition errors for proper nouns.

Original post →

More from coding & agent

coding & agent channel →