Transcripts

show.fm can turn an episode's audio into an accurate, editable transcript. Review it, fix speaker names, publish it to your RSS feed so podcast apps can show it to listeners, or export it for use anywhere else.

Before you start

• AI transcription is available on Creator plans and up. On the Free plan the transcript card shows an Upgrade to unlock button instead.
• On an externally hosted show, upload the recording on the Media tab's Your recording card first. show.fm makes the transcript from it and never serves it to listeners. The recording is deleted 30 days after the episode goes live; the transcript stays, and the editor says so when the player is gone. See Externally Hosted Shows.
• Transcription works with MP3 and M4A audio, up to six hours per episode. Every run first prepares the audio through Audio Processing for the most accurate results, so a typical episode takes several minutes end to end. WAV uploads are converted to MP3 automatically and can be transcribed as soon as conversion completes. Very large MP3 files (over 100 MB) depend on the preparation step: if preparation is temporarily unavailable the transcription can fail with a size message, and retrying later succeeds.

AI credits

Transcription uses AI credits from your plan's monthly allowance. One credit covers roughly one minute of audio, and the card shows an estimate before you start.

Plan	AI credits per month
Free	None
Creator	120
Studio	300
Production House	1,200
Enterprise	Unlimited

If a run would exceed your remaining credits, it doesn't start; you'll see how many credits it needs versus how many remain. See Billing and Plans.

Transcribe automatically

Turn on Automatically transcribe new uploads in your Podcast Settings (the AI card on the General tab). New uploads are then transcribed for you once the audio is in a publishable format (WAV uploads start transcribing automatically after conversion), using about 1 credit per minute of audio. The six hour limit applies here too: a longer upload is not transcribed at all, and you get a notification explaining the limit. Uploads that would exceed your remaining credits are also skipped, with a notification so you know.

Re-uploading replaces edits

With auto-transcribe on, re-uploading an episode's audio re-transcribes it and replaces any transcript edits you've made.

Transcribe one episode

1. Open the episode and go to the Media tab.
2. In the Transcript card, pick the Spoken language, or leave it on Auto-detect.
3. Check the credit estimate, then click Generate transcript.
4. A dialog shows roughly how long the run will take and confirms you can leave the page while it works. Click Start transcription.

A progress bar tracks the stages (preparing, transcribing, identifying speakers, saving). Identifying speakers is the longest step on longer episodes. You can leave the page; transcription keeps running in the background and you'll get a notification when it's done.

When it finishes, the card shows the word count, duration, and language, plus badges for the transcript's state (AI draft, Edited, Published). Click Review transcript to open it.

If the episode already has a transcript imported from another host, generating an AI draft keeps that transcript in the feed. It is replaced when you publish the new draft.

The episode list shows a Transcribed pill when an episode has a saved transcript or an imported transcript linked to its feed. A saved draft does not need to be published to show this pill. Co-hosts do not see it when transcript access is hidden.

The transcript surface

The transcript opens inside the Media tab rather than as a page of its own. Media stays highlighted, a Media back link sits above the transcript, and the Episode chat panel stays where it is, so you can talk about a passage while you read it.

If no transcript exists yet, you'll see No transcript yet with a Go to Media button (generation lives on the Media tab). On plans without AI transcription it shows Upgrade to unlock instead.

The transcript surface shows your transcript as paragraphs grouped by speaker, alongside a sticky audio player. Click a paragraph to jump the player to that moment.

• Speakers: the AI labels each voice. Use the speaker chips above the transcript to rename speakers, match them to people on the episode, add a speaker, or remove one. You can also change the speaker on individual paragraphs.
• Edit text: use the pencil beside a paragraph and type. Apply changes closes that paragraph edit. Click Save in the black bar, or press Ctrl+S / Cmd+S, to persist all text and speaker changes, including an edit still open on screen. Discard restores the last saved transcript. The page asks what to do before you leave with edits.
• Save without audio: the same save bar works if the recording has expired or is unavailable. On a phone it remains available when you switch to Promo assets.
• Find and replace: search the transcript, with case-sensitive and whole-word options, and replace every match at once. Great for a name the AI keeps mishearing.

Status chips above the transcript show Published and Edited states. If you edit after publishing, a notice reminds you the published version is older than your edits.

Publish the transcript to your feed

1. Above the transcript, click Publish to feed (after the first publish, the button reads Publish update).
2. Confirm in the dialog.

The saved transcript is rendered as a WebVTT file at a public URL and linked from your RSS feed, so podcast apps can show it to listeners. Publishing again creates a new public file address for the same episode. Podcast apps may take time to show the update. Unsaved changes are not included; save first, then publish.

Publishing also puts the transcript on the episode's public listen page. When both the audio and transcript are hosted by show.fm, Read the transcript on the Transcript card, or Transcript on the player, opens a full-screen reading view with the player pinned on top. Listeners can search, select a line to jump to that point, and follow along. The current word is highlighted when word timings are available.

Speaker names you set here are used there. A speaker whose name matches someone credited on the episode shows that person's photo beside their lines, or their initials when they have no photo.

The Transcript card below the description shows the first lines and offers Open transcript file, including when JavaScript is off or the audio is unavailable. Follow-along needs both the audio and transcript to be hosted by show.fm. See Smart links for the listening controls.

Word-by-word follow-along needs a fresh publish

Word timings are written into the transcript file when you publish. A transcript published before this feature existed, or one imported from another host, still follows along a line at a time. Publishing the transcript again picks up the word timings.

To remove it, open the transcript's overflow menu and choose Unpublish from feed…. The public link is removed from your feed and its published addresses are withdrawn as delivery updates propagate. Apps may keep a cached copy for a while. You can publish again at any time.

Re-run the transcription

Choose Re-run transcription… from the overflow menu to replace the transcript with a fresh AI pass. This replaces everything, including every edit, split, and speaker assignment, and spends AI credits again. When the new transcript finishes saving, the previous version is unpublished. Review the new transcript and publish it when it is ready. The confirmation button is Replace transcript.

Review a transcript after replacing audio

Replacing the episode audio removes the old transcript from publication. Your saved text remains available, with a warning that its words or timings may no longer match. This also applies if you remove the recording before uploading a new one.

1. Open Review transcript from the episode's Media tab.
2. Check the text and timings against the new recording, or choose Re-run transcription… to generate a new draft.
3. Save any edits, then choose Publish to feed when the transcript is ready.

Saving edits alone does not update the public transcript. A podcast app may show its own generated transcript instead of the file you published. See Replace episode files for download and refresh limits.

Export the transcript

The Export menu downloads the transcript in five formats:

• SRT: subtitle file for video editors and players
• VTT: web-standard captions
• Plain text: just the words
• Markdown: a .md document with the episode title and paragraphs grouped by speaker
• JSON: structured data with timings and speakers

Promo assets

The Promo assets panel beside the transcript generates promotional content from it: three title ideas, a short episode summary, a tweet, a LinkedIn post, and suggested tags. Click Generate content to create them, or Regenerate for a fresh set. Generation uses AI credits from your monthly allowance, billed on actual usage; a typical episode costs a few credits.

What's Next?

• Publish the episode
• Share and embed the episode
• Check your plan's AI allowance