Podcast transcription from your episode file

How It Works

Upload Your Recording

Choose a supported episode file, or record in the browser if you are capturing a short pickup. Speaker detection starts enabled with a two-speaker hint, which matches a host-and-guest show. Open settings to raise the count, turn detection off, or add names such as a co-host before you submit. Select the spoken language or leave auto-detect on, then read the duration and credit estimate. If login interrupts a draft, follow the file-restoration prompt. A filename alone cannot restore an audio file after a refresh.

Review the Result

Progress, queue state, and the finished podcast transcript stay on this page. Play the episode back, search a phrase, and rename speaker labels in the speaker panel. Treat those labels as groups inside this recording, not verified identities. Intro music, ads, and two people talking at once will still need a human pass. If a task fails, read the error before starting another attempt. Refreshing a processing task restores its status from the server.

Export Your Text or Captions

Save corrections first. Download DOCX for an episode-page draft, TXT for a plain archive, or SRT and VTT if you are publishing a video version of the same show. Exports use the saved transcript. An optional summary is a later action on the same record when your plan allows it; a failed summary leaves the transcript in place. The task also appears in My Recordings next to work started in the general transcription workspace. Reopening it does not start a second transcription or charge the base fee again.

Prepare a Transcript for Your Episode Page

A usable podcast transcript is the spoken record: who spoke, in what order, with timestamps you can check against the file. Show notes are a different artifact. They name the topics, guests, and links a listener needs after the episode. Mixing the two produces a page that is hard to quote and hard to scan.

After the run finishes, replace Speaker 1 and Speaker 2 with the host and guest names you already use on the show. Confirm proper names, numbers, and sponsor reads against your rundown. Keep filler and false starts if you are archiving; clean them only when you are publishing a reading copy. Do not paste an unreviewed transcript under the player and call it show notes.

This page supports that review for episodes you recorded or received as a file. It is not a podcast host, not a directory, and not a replacement for permission from every speaker on the tape.

See an Example

Harbor Ledger is synthetic owned practice input written for this page. The MP3 is not a transcription-provider output, and it is not taken from a real show. Save the file, then drop it into the upload box above if you want to walk the live path. Playback of the sample never submits a task.

Each row is the authored script for that beat, timed from the committed MP3. A short sine tone stands in for intro music so you can hear what a theme sting does to a first pass. The last beat stacks both voices on purpose. Check the returned wording, speaker grouping, and clocks against this file. Confirm the proper name Harbor Ledger, and keep the spoken line about show notes separate from any summary you write later.

1. 00:00Host
   Welcome to Harbor Ledger. After the theme, we will talk about last week's harbor delay.
2. 00:05Theme stingTheme
   A 0.4 second sine tone stands in for intro music. It is not a licensed theme.
3. 00:06Guest
   The Tuesday barge left two hours late. The crate count stayed the same.
4. 00:10Host
   So the delay was time, not cargo?
5. 00:13Guest
   Yes. We still need the dock notes before we publish show notes.
6. 00:17Host / GuestOverlap
   HostSorry, go on.

   GuestKeep the transcript separate from the show notes.

Supported Files and Plan Limits

This page accepts MP3, WAV, M4A, MP4, MOV, FLAC, OGG, WEBM, AAC files, up to 1 GB and 60 minutes each. Browser recording requires microphone permission and browser support. The bytes have to be readable audio or video with a soundtrack. Changing an extension does not convert a file.

Billing follows duration, not file weight: 0.5 credits through 30 seconds, then one credit for every started minute. The API re-checks size, length, and remaining credits on submit. Read the estimate above and the plans and usage limits page before you start a full episode. There is no unlimited free pool and no local-only mode on this route.

Theme music, ads, compressed MP3s, and overlapping hosts make words and speaker groups less reliable. Speaker count is a hint, not a guarantee. If you later request a summary or translation, those steps use the same saved transcript and the existing plan rules. They are not generated automatically for search-engine copy.

Frequently Asked Questions

Can I transcribe an episode from a podcast URL?

No. This page does not fetch a Spotify, Apple Podcasts, YouTube, or RSS URL. Download the episode file you own or have permission to process, then upload it here. The general transcription workspace can transcribe some direct media URLs. It still does not log into a podcast app or subscribe to a feed.

Are show notes the same as a transcript?

No. A transcript is the spoken record, including asides, sponsor reads, and false starts unless you edit them. Show notes are an editorial summary with names, topics, and links. If you need a shorter recap after the transcript succeeds, use the existing summary action on the same task when your plan includes it. A failed summary does not remove the transcript.

Does intro music break podcast transcription?

It often confuses the first seconds. The practice clip includes a short tone so you can see that effect. Trim a long theme in your editor if the opening is more music than speech, then upload the spoken body. Lyrics are not a promised result.

What happens when two people talk at once?

Overlap is marked in the practice script on purpose. Automatic speaker groups can split, merge, or drop words when voices sit on top of each other. Replay that passage and keep an unclear word marked instead of guessing before you publish.

Is this free, and does it run on my machine?

Neither as a default. New accounts start with 5 free minutes. After that, work is billed in credits against your plan. Audio is uploaded and processed in the cloud. Choosing a file does not start a run until you sign in and submit.

Who can open my episode after I leave?

Task and download requests require the owning account. Open the same record from My Recordings or from this page's task link. Other users cannot read the file or transcript by guessing the task id. Create a sharing link only when you intend to, and revoke it when access should end.