Interview transcription: formats, examples and how to do it fast

This is the hub for our interviews and transcripts guides. When the transcript is done, from transcript to story covers finding, quoting and checking the material.

The three interview transcription formats

The format is a decision, not a detail. Choose it before you start, write it at the top of the transcript, and use the same one for every interview in a project.

Interview transcription example in each format

The interview below is a fictional example. The speaker, the clinic and every detail are invented.

Verbatim

[00:04:12] INTERVIEWER: So when did you first notice the wait times
getting longer?

[00:04:16] DANA REYES: Um, I mean, I'd say, uh, it was, it was probably
around March? Because that's when, that's when Dr. Patel left, and we
didn't, we didn't have anyone to cover her patients for, like, [pause]
three months. [laughs] Which was, yeah.

Intelligent verbatim

[00:04:12] INTERVIEWER: When did you first notice the wait times getting
longer?

[00:04:16] DANA REYES: I'd say it was probably around March, because
that's when Dr. Patel left, and we didn't have anyone to cover her
patients for three months.

Edited

INTERVIEWER: When did wait times start getting longer?

REYES: Around March, when Dr. Patel left. We had no one to cover her
patients for three months.

Look at what changed. Intelligent verbatim keeps "I'd say" and "probably," which carry the speaker's uncertainty. The edited version drops them and reads as more certain than she was. That is fine for a Q&A labeled "edited for length and clarity," but you should not quote the edited line as her exact words.

Interview transcription format: conventions and template

Conventions

Template (copy and paste)

# Interview transcript: [Interviewee full name]

- Interviewee: [Full name], [title, organization]
- Interviewer: [Name]
- Date and time: [Mon. D, YYYY, h:mm a.m./p.m., time zone]
- Location or medium: [in person, address / phone / video platform]
- Recording file: [file name], length [hh:mm:ss]
- Consent: [consent given on the recording at 00:00:15 / written form on file]
- Ground rules: [on the record / background / any off-the-record sections]
- Format: [verbatim / intelligent verbatim / edited]
- Transcribed by: [name or tool], checked by [name] on [date]

---

[00:00:00] INTERVIEWER: [...]

[00:00:00] [SURNAME]: [...]

---

Spellings confirmed: [names, places, organizations]
Follow-ups: [documents promised, questions to ask again]

The header matters more than it looks. Six months later, "consent given on the recording at 00:00:15" and "checked by" are what let you, your editor or a review board trust the file.

How to transcribe an interview: manual, automatic or hybrid

The hybrid method, step by step:

1. Record well. Put the recorder close to the speaker, away from fans and cafe noise. Record a backup on a second device.
2. Run the automatic draft. Add names and technical terms to the tool's custom vocabulary if it has one.
3. Fix speaker labels first; automatic speaker separation often swaps people who sound alike.
4. Correct every name, number and technical term by listening to those spots.
5. Listen through every passage you might quote, word for word, at normal speed.
6. Mark what you could not hear as [inaudible hh:mm:ss] instead of accepting the tool's guess.
7. Fill in the header and save the transcript with the audio file name in its title.

Step 5 is the one people skip and the one that matters. Automatic transcripts can be wrong in ways that look right.

Interview transcription software compared

Prices and claims below come from each vendor's own pages as of November 2026. Plans change often; check before you buy.

How to choose:

• Fastest free start: Otter's or Rev's free tier for a few short interviews.
• Publication-grade accuracy without doing it yourself: a human service such as Rev's.
• You edit audio or video too: Descript, which lets you edit the recording by editing the text.
• A newsroom team: Trint, which is built around newsroom collaboration.
• Sensitive sources: a local Whisper app, or manual transcription. Nothing goes to a third party.

A warning about automatic accuracy

Automatic transcripts fail in two ways. The obvious one is garbled words, which you can see. The dangerous one is fluent text the speaker never said. A 2024 study of Whisper by Allison Koenecke and colleagues found that roughly 1% of the audio transcriptions they tested contained entire hallucinated phrases or sentences that did not exist in the audio, and that longer pauses made them more likely. [8] The researchers noted the model has since improved. The lesson applies to every tool: before you quote anything, listen to it.

Consent and privacy: recording laws and storing transcripts

Recording consent laws

This is general information, not legal advice. The Reporters Committee for Freedom of the Press publishes a state-by-state recording guide and runs a legal hotline for journalists. [9]

• Federal law requires the consent of at least one party to record in-person, phone or electronic conversations. If you are in the conversation, your consent counts.
• One-party consent states follow that rule. Most states are one-party.
• All-party consent states require everyone in the conversation to agree. According to the Reporters Committee, the states that primarily require all-party consent are California, Delaware, Florida, Illinois, Maryland, Massachusetts, Michigan (at least for third-party recordings), Montana, New Hampshire, Pennsylvania and Washington. [9]
• Mixed rules: Missouri and Oregon require all-party consent for in-person conversations but one-party for phone calls; Connecticut and Nevada are the reverse; Hawaii and Maine require all-party consent only in particularly private places. [9]
• Across state lines, the Reporters Committee advises journalists to assume the stricter state's law applies. [9]

The habit that covers almost every case: start the recording, then ask, "Is it OK if I record this conversation?" and get the yes on tape. Researchers working with human subjects also need the consent their ethics board approved, which usually covers storage and sharing of transcripts as well.

AI meeting notetakers raise a newer issue. In August 2025, a proposed class action filed in federal court in California alleged that Otter's notetaker recorded meeting participants who had not agreed to it and used the recordings for AI training; the claims had not been decided in the coverage we found. [10] If a bot joins your call, tell everyone and get their consent.

Storing transcripts safely

The Freedom of the Press Foundation's review of transcription tools recommends: [4]

• Use a long, unique password and turn on two-factor authentication where offered.
• Download your files and delete them from the service when you no longer need them there; set a retention rule.
• For highly sensitive recordings, transcribe locally or manually.
• Check your organization's or contract's retention requirements, including any legal hold, before deleting.

Also: keep the audio. A transcript is only as trustworthy as the recording you can check it against.

After transcription: find quotes and check them in WriteOrbit

WriteOrbit does not transcribe audio or video. Use one of the tools above to get the transcript, then bring it into WriteOrbit for the next step: finding what matters, quoting it accurately and writing beside it.

1. Save the transcript as a PDF and upload it to the Library. Name it with the interviewee and date, for example 2026-11-04_Reyes_intelligent-verbatim.pdf. The free plan holds 10 PDFs up to 10 MB each; Plus allows unlimited PDFs up to 50 MB.
2. Use Ask on that transcript. Ask works on one PDF at a time, and its answers carry page numbers: "Where does Reyes talk about staffing?" or "What dates does she give for Dr. Patel leaving?" Open the cited page, find the timestamp, and listen to that spot in the audio before you quote it.
3. Or open the Word file as a document. WriteOrbit opens .docx and keeps its styles, so you can comment on a transcript, highlight quotes and share it with an editor or co-researcher.
4. Write the story or report beside it. The AI Chat panel next to your draft can use your Library, your document and the web as sources, which helps you look across several sources at once; treat its answers as leads and confirm each one in the transcript itself.
5. Run AI review on the draft. Claim checks and proofreading appear as comments on the exact sentences.

From transcript to story walks through this workflow in detail, including AP quote rules and a thematic analysis method for researchers.

FAQ

What is an interview transcription?

An interview transcription is a written record of a recorded interview, with each speaker labeled and, usually, timestamps that point back to the audio. It comes in three formats: verbatim, which keeps every word and sound; intelligent verbatim, which drops filler but keeps meaning; and edited, which also tidies grammar for publication.

How long does it take to transcribe a one-hour interview?

By hand, expect several hours of work for each hour of audio, depending on your typing speed, audio quality and how many people talk at once. Automatic tools return a draft in minutes. Most reporters and researchers now use a hybrid: an automatic draft, then a human pass to correct names, numbers and every passage they plan to quote.

Which transcription format should I use?

Use verbatim for legal work and research that analyzes how people speak, including pauses and false starts. Use intelligent verbatim for most journalism and qualitative research, because it is readable and keeps the meaning. Use edited transcripts only for published Q&As, and tell readers the text was edited for length and clarity.

Is it legal to record an interview?

Federal law and most states allow recording when one party to the conversation consents, so you can record your own interview. About a dozen states, including California, Florida, Illinois, Maryland, Massachusetts, Pennsylvania and Washington, generally require everyone's consent. Across state lines, assume the stricter law applies. Asking on the recording settles it.

Can WriteOrbit transcribe audio?

No. WriteOrbit does not transcribe audio or video. Use a transcription tool or service first, then bring the transcript in: upload it as a PDF to the Library and use Ask to find passages, with page numbers, or open the Word file as a document. You then write the story or report beside it.

Next steps

A transcript is raw material. Read from transcript to story next for how to choose quotes, attribute them in AP style and check them. For where transcription fits in a full reporting workflow, see AI tools for journalists and how to write a news article.

Sources

1. Otter.ai, Pricing. https://otter.ai/pricing
2. Otter.ai, Privacy Policy (last updated June 16, 2026). https://otter.ai/privacy-policy
3. Rev, Pricing. https://www.rev.com/pricing
4. Martin Shelton and Yael Grauer, "How secure are journalists' favorite transcription tools?" Freedom of the Press Foundation (updated Oct. 1, 2026). https://freedom.press/digisec/blog/how-secure-are-journalists-favorite-transcription-tools/
5. Descript, Pricing. https://www.descript.com/pricing
6. Descript Help Center, "Data privacy." https://help.descript.com/account-and-app-settings/data-privacy
7. OpenAI, Whisper (GitHub repository). https://github.com/openai/whisper
8. Allison Koenecke et al., "Careless Whisper: Speech-to-Text Hallucination Harms," ACM FAccT 2024. https://arxiv.org/abs/2402.08021
9. Reporters Committee for Freedom of the Press, "Reporter's Recording Guide: Introduction." https://www.rcfp.org/introduction-to-reporters-recording-guide/
10. NPR via WUNC, "Class action suit claims Otter AI secretly records private work conversations," Aug. 15, 2025. https://www.wunc.org/2025-08-15/class-action-suit-claims-otter-ai-secretly-records-private-work-conversations