A raw transcript is not a finished product. It is a long block of text with no speaker names, plenty of "um" and "you know", and a few misheard names. This guide covers what to do next: how to clean the transcript, and how to turn it into show notes, chapter markers, quotes and a blog recap.

Key Takeaways

Edit in passes. Fix speakers first, then names and jargon, then filler words, then timestamps.
Keep one clean "master" transcript. Build show notes, chapters, quotes and a recap from it.
Check every quote against the audio before you publish it.
Publish the transcript and the recap as different things. The recap is for readers. The transcript is the record.

Step 1: Save a Copy of the Raw Transcript

Before you edit anything, export the raw version and keep it. If you used timestamps, export a timestamped copy too (TXT with timestamps, or SRT). You will need those times later for chapter markers. Then make a second copy to edit. If you make a mistake, you can always go back.

Step 2: Add Speaker Labels

Speech-to-text output usually does not tell you who is talking. For a two-person interview, this is quick to fix. Read the first few minutes, find where the host stops and the guest starts, and add labels like HOST: and GUEST (Maria):.

Some tips that save time:

Questions usually come from the host. Long answers usually come from the guest.
Start a new paragraph every time the speaker changes, even for one-word replies like "Right."
For panels of three or more, write a short list of speakers and how each one sounds or talks. Use it as you go.
Where two people talk over each other, the text may be mixed up. Mark it with [crosstalk] rather than guessing.

Step 3: Fix Names, Jargon and Numbers

This is where most real errors hide. Speech models guess at words they have not seen often. A guest named "Siobhan" can come out as "Shivon". A product called "Kubernetes" can come out as three random words.

Make a short word list before you start. Include the guest's name, their company, any products, places and technical terms from the episode. Then search the transcript for close misspellings and fix them all at once with find-and-replace. Also check every number: years, prices, percentages and dates. A wrong number in a quote is worse than a missing one.

Step 4: Decide How Much Filler to Remove

There are two common styles:

Verbatim: keep everything, including "um", false starts and repeats. Useful for research or legal records.
Clean read: remove filler words and false starts, but keep the speaker's meaning and voice. This is what most podcasts publish.

For a clean read, delete "um", "uh", "like" (when it is filler) and repeated words such as "I, I, I think". Do not rewrite sentences to sound smarter. If the guest said "it was kind of a mess", leave it. Pick one style and use it for every episode so your transcripts look consistent.

Step 5: Tidy the Timestamps

A timestamp on every line is hard to read. For a published transcript, keep one timestamp at the start of each speaker turn, or one every few minutes. Use a single format, such as [12:40]. Keep the full timestamped export for yourself. You will use it in the next step.

If you still need a transcript in the first place, our podcast transcription tool takes an episode file and gives you text with optional timestamps. The free version processes files locally in your browser, with no upload, and supports 100+ languages. If the episode is a video file, video to text takes that file and can export SRT or VTT from the same transcript.

Turn the Clean Transcript Into Show Notes

Show notes help a listener decide whether to press play. Keep them short. A simple structure works well:

One or two sentences on what the episode is about and who the guest is.
Three to five bullet points on the main topics, in the order they come up.
Links mentioned in the episode: books, tools, the guest's website.
Chapter markers (see below).

Example: "Maria runs a small bakery in Lisbon. We talk about pricing a new menu, hiring a first employee, and what she would do differently." That is better than a vague line like "An inspiring talk about business."

Create Chapter Markers

Chapters let listeners jump to the part they care about. Go through your timestamped copy and note where the topic changes. Aim for one chapter every 5 to 10 minutes. Give each one a plain, specific title.

00:00 Intro
03:15 How Maria found her first location
11:40 Pricing the new menu
24:05 Hiring the first employee
37:30 What she would do differently

Many podcast hosts and apps read chapters from timestamps in the episode description. Check your host's help pages for the exact format it needs, such as whether the first chapter must start at 00:00.

Pull Out Quotes

Look for short lines that make sense on their own. Good quotes are specific and one or two sentences long. Copy them from the clean transcript, then play that part of the audio and check the wording. Make sure the quote still means what the speaker meant once it is out of context. Add the speaker's name and the timestamp so you can find it again.

Write a Blog Recap

A recap is not the transcript pasted into a blog post. It is a new, shorter article built from the episode. A good recap:

Has its own title that says what a reader will learn.
Covers the three or four best ideas, each under its own heading.
Uses one or two checked quotes.
Links to the episode and to the full transcript.

You can use an AI writing tool to draft the recap from your clean transcript. If you do, read every line against the transcript. Remove anything the guest did not actually say.

Checklist

Raw transcript and timestamped copy saved
Speaker labels added; crosstalk marked
Names, products, jargon and numbers checked
Filler words removed (clean read) or kept (verbatim), consistently
Timestamps reduced to one format
Show notes written, with links
Chapter markers added in your host's format
Quotes checked against the audio
Blog recap drafted and fact-checked

FAQ

Should I publish the full transcript or just a summary?

Publish both if you can, but keep them separate. The full transcript is useful for accessibility: people who are deaf or hard of hearing, or who prefer reading, can follow the whole conversation. It also lets anyone check exactly what was said. A summary or recap is easier to read and share, but it leaves things out, so it cannot replace the transcript for those readers. A practical setup is an episode page with the show notes and chapters at the top, the recap below, and the full clean-read transcript at the bottom or on a linked page. If time is short, publish the transcript first. A recap can come later.

What to Do After You Have a Podcast Transcript: Editing, Show Notes and Chapters