Getting the words right is only half of a good transcript. The rest is presentation: who is speaking, what the listener can hear besides speech, and how the text is broken up so a human can follow it. Two transcripts with identical wording can be pleasant or exhausting to read depending on these choices. This guide covers the decisions that matter most and ends with a simple style guide you can copy for your own show.

Verbatim or clean verbatim?

The first decision is how faithfully to capture speech. Full verbatim records everything: every “um,” false start, repeated word and stammer. Clean verbatim keeps the meaning and voice but removes filler and accidental repetition.

For most podcasts, clean verbatim is the better fit. It is easier to read and still honest. Reserve full verbatim for cases where the hesitations carry meaning, such as an investigative interview, a legal matter or an oral history, where how something was said is part of the story.

Whichever you choose, apply it consistently, and never rewrite what a person said to make them sound smarter or to change their meaning. Removing a filler word is editing for readability. Changing a word is altering the record. Keep slang and distinctive phrasing, because that is how people sound. If a speaker’s grammar is unusual, leave it unless it would confuse the reader.

Speaker names and labels

Readers need to know who is talking, and a screen reader user cannot glance at a face or voice to work it out. Label each speaker turn clearly.

• Use the full name on first appearance, and a consistent short form afterward.
• Put the label at the start of each turn, in bold or on its own line, so it is easy to spot.
• Use the same label for the same person throughout the transcript.
• When a speaker’s name is unknown, use a descriptive label such as “Caller” or “Audience member” and keep it consistent.

Start a new paragraph whenever the speaker changes. If one person talks for a long time, split their turn into paragraphs and repeat the label only if the page breaks make it unclear who is speaking. Make sure the label is real text, not an image or an icon, so assistive technology reads it.

Describing sounds and music

Some audio carries information that words do not. Laughter, a door slamming, a long silence, theme music and a change of location all shape the listening experience. People who cannot hear the audio should not lose that context.

Put non-speech sounds in square brackets, in plain language, and describe only what matters:

• [laughter]
• [phone rings]
• [upbeat theme music fades in]
• [both speakers pause for several seconds]

Be specific rather than poetic. “[tense music]” helps; “[a haunting melody drifts like fog]” does not, unless the show itself is deliberately stylized. Describe the sound’s purpose, not every noise in the room. If the background hum of a café never affects the conversation, leave it out. If a sound effect is the punchline, include it.

Music deserves a short note on when it starts and stops, particularly if it is a recurring theme. If lyrics are meaningful and you have permission, include them. Otherwise name the song or describe the genre briefly.

Describe what changes the meaning of the moment, not everything that makes a noise.

Timestamps

Timestamps help readers find a spot in the audio, and they help you when you need to check a quote. You do not need one on every line. Placing them at paragraph breaks or topic changes, or every few minutes, is usually enough. Use one format throughout, such as minutes and seconds in square brackets at the start of a paragraph. Remember that they add clutter, so a transcript with a timestamp on every sentence becomes tiring to read and noisy for a screen reader. If your player lets people jump to a time, timestamps can double as navigation, but that is an extra, not a requirement.

Paragraphing and punctuation

Speech does not come with paragraphs, so you have to supply them. Break the text where the topic shifts, where the speaker changes or where a listener would naturally take a breath. Aim for paragraphs short enough to scan on a phone. Use punctuation to mirror meaning: question marks for questions, dashes for interrupted thoughts, and ellipses sparingly for trailing off. Avoid long, unpunctuated blocks, which are hard for everyone and especially tiring for readers with cognitive or attention difficulties.

Proofread names and jargon

This is the step that separates a decent transcript from a trustworthy one. Names, places, brands, medical or legal terms, and acronyms are where errors hide, and they are also what readers search for. Make a list before you start, ask guests how to spell their names, and check unfamiliar terms against a reliable source. Then search the finished text for each term, to catch variant spellings.

A simple style guide

Consistency is easier when the rules are written down. A one-page guide is plenty. Include these items:

1. Verbatim style: clean or full, and what counts as a filler word.
2. Speaker format: full name first time, short form after, label style.
3. Sound notes: square brackets, lowercase, present tense, what to include.
4. Timestamp rule: format and how often.
5. Numbers: when to spell out, when to use digits.
6. Spelling list: names, brands and terms, with preferred capitalization.
7. Language and tone: how to handle profanity, slang and unclear audio.

For unclear audio, choose a marker such as [inaudible] and use it sparingly, never as a way to skip hard passages.

Review the guide after your first few episodes. You will discover cases it does not cover, such as how to handle a guest who switches languages mid-sentence or a recording with overlapping voices. Add a rule each time, and share the page with anyone who edits for you, so the whole archive reads as if one careful person wrote it.

Final thoughts

A transcript is a reading experience in its own right. Decide how faithfully you will capture speech, label every speaker, describe sounds that matter, and break the text into manageable pieces. Then write your decisions down so every episode follows the same rules. Do that, and readers will trust your text as much as they trust your voice.