Notes on transcription

Short, practical reads on getting clean transcripts out of messy audio.

How accurate is AI transcription in 2026?

Word error rates by model, and when to still send a human pass.

6 min read

SRT vs VTT: picking the right subtitle format

The three differences that actually matter for your editor.

4 min read

Transcribing multi-speaker interviews without losing the thread

Speaker labels, overlapping talk, and how to clean it up fast.

7 min read

The 30-minute rule: why shorter uploads transcribe faster

What actually happens to a file the moment you drop it in.

3 min read

Timestamps are the whole point

Why every line should be a seek point, not just a caption.

5 min read

A field guide to filler words

When to keep "um" and "like" — and when to cut them.

4 min read

Recording audio that transcribes cleanly

Mic placement, room noise, and the codec choices that help most.

8 min read

Exporting to PDF vs DOCX vs plain text

Which format survives a client email thread intact.

3 min read

How language detection works across 100+ languages

What the model looks at in the first ten seconds of audio.

6 min read

Podcast editors: turn a transcript into a rough cut

Using timestamps to find your best ninety seconds, fast.

5 min read

Lecture capture: transcribing a whole semester

Folder structures that survive fourteen weeks of uploads.

6 min read

Why longer files queue — and how to skip it

Priority processing on the Basic and Advanced plans, explained honestly.

4 min read