RetentaRetenta

Evidence record · captured 2026-10-10

OpenAI Audio Transcription

A repeatable CLI workflow for OpenAI speech transcription, optional speaker diarization, known-speaker hints, chunking, and organized output validation.

Official publisherUnverifiedSource reviewedBehavior not tested
Source review

Editorial assessment

This is a source-only assessment. It does not claim runtime or security testing.

What it does

The source chooses a fast text model by default, switches to a diarization model only when speaker labels are requested, supports limited known-speaker references, and requires output-quality review.

Best for

  • Turning interviews or meetings into text
  • Speaker-labeled transcription with optional reference clips
  • Repeatable batch transcription saved to explicit output directories

Strengths in the source

  • Separates the fast default from the more specialized diarization path.
  • Keeps API keys out of chat and documents deterministic CLI commands.
  • Calls for validation of speaker labels and segment boundaries rather than treating API output as final.

Limitations

  • Requires sending audio to OpenAI and incurring API usage.
  • Known-speaker hints are capped at four in the documented example path.
  • The workflow is file-oriented, not a live-streaming transcription design.
  • The pinned publisher repository is deprecated in favor of OpenAI's plugins repository, so confirm the current distribution path before adopting it.

When not to use it

  • Audio must remain fully offline or on premises
  • The task requires real-time captions rather than file transcription

Reviewed by Retenta editorial on 2026-10-10; scope: source only.

Files included in this review
  • skills/.curated/transcribe/SKILL.md
  • README.md
  • skills/.curated/transcribe/LICENSE.txt

Evidence and boundaries

Execution · not tested

Retenta reviewed the pinned source but did not install or execute this skill.

Security · not tested

No sandbox run, dependency audit, or independent security assessment is claimed.

Revision
49f948faa9258a0c61caceaf225e179651397431
License evidence
Apache-2.0The pinned skill directory includes an Apache License, Version 2.0 LICENSE.txt.
Last source change
2026-01-31
Independent verification
The source is controlled by OpenAI; Retenta has not independently verified its runtime behavior or security.
Upstream repository
Open publisher repository

Compatibility

Codex

documented

OpenAI's repository documents curated skills for Codex and installation through the built-in skill installer.

Requirements and access

  • An OPENAI_API_KEY set locally
  • Python and the OpenAI package
  • One or more audio or video inputs

Referenced tools

transcribe_diarize.pyOpenAI transcription API

Install from the source

The pinned README documents name-based installation for curated skills and says to restart Codex. It also marks this repository deprecated in favor of OpenAI's plugins repository, so confirm the current path before adopting it.

View the SKILL.md at its source Verify the publisher's install documentation

More Documents, data and media skills →

Codex skill installer

$skill-installer transcribe

Thyme RAG Capsules

Make this skill’s state durable, governed, and portable.

A Retenta capsule lets agents recall prior work, append provenance-stamped progress, and carry the memory to the next run.

Create durable agent memoryUnderstand the evidence labels