This is just a little utility for automatically transcribing the contents of a personal audio recorder. I use one for managing the noise in my ADHD-addled head, and the Whisper speech-to-text engine makes the transcription very easy.
Go to file
Elf M. Sternberg e95ec93509 Moved to PDM 2022-11-08 11:34:52 -08:00
.gitignore Moved to PDM 2022-11-08 11:34:52 -08:00
.pdm.toml Moved to PDM 2022-11-08 11:34:52 -08:00 A little too to assist with TCD transcription. 2022-10-04 10:23:17 -07:00 Revised styling 2022-10-15 06:52:36 -07:00
pdm.lock Moved to PDM 2022-11-08 11:34:52 -08:00
pyproject.toml Moved to PDM 2022-11-08 11:34:52 -08:00 Wrap transcription attempt in a try/except 2022-11-08 10:58:41 -08:00

Language: Python Topic: Utility

TCDump: Thought Capture Device Transcriber

My current TCD
I have ADHD and I use a small tape recorder to capture all of the ephemeral ideas and incidents that happen around me. The OpenAI project recently released [Whisper](, a speech-to-text engine that has surprisingly good recognition even with it's smallest model.

This script is extremely personal and specific: it effectively transcribes and dumps the contents of my tape recorder as text to stdout. That's out. It's extremely useful to me, especially in conjunction with the Unix tee program.

If you have a tape recorder that presents as a USB device, you could probably just make this work for you by:

  • editing the path to the repository of recordings
  • the regular expression describing the file name convention
  • the strptime format in converttcdfilename that turns the filename into a human-readable date.

Note that "Could not find the repo" is an error condition (exit status 1), but "No files found in repo" or "No files found meeting your criteria" or "No new files found since last transcription" are all exit status 0, matching the Unix tradition.


TCDump is Copyright Elf M. Sternberg (c) 2019, and licensed with the Mozilla Public License vers. 2.0. A copy of the license file is included in this folder.