Which file to upload
You usually have two choices: the finished episode or the raw recording. The finished MP3 that goes to your host is ready to upload and matches what listeners hear. A video podcast gives you an MP4 or MOV, and the audio is extracted automatically.
If your intro or interludes have music under the voices, consider uploading the voice-only edit instead. A music bed under speech lowers accuracy and can produce stray words during instrumental parts. Remote recording tools such as Riverside, Zencastr and SquadCast can give you one track per guest. Transcribe the mixed episode for publishing, and keep the separate tracks for editing.
Recording choices that show up in the transcript
Good podcast audio and a good transcript come from the same habits. Dynamic microphones close to the mouth reject room sound. Headphones on every host and guest stop echo. Leaving a short pause instead of talking over each other makes both the edit and the transcript cleaner.
Spell out unusual names, product names and acronyms when you first say them, or at least have them written down. Speech recognition guesses these, and you will want to check them before publishing.
If you record remotely, ask guests to record a local copy on their side as well. Internet dropouts cause gaps and garbled words in the call recording, and a local track lets you patch those parts in the edit before you transcribe the finished episode. Keep the room quiet during recording too. Fans, keyboards and notifications all end up in the audio.
From transcript to show notes and episode page
After the transcript is ready, play it back with the synced player and fix names and terms in place. Click any line to jump to that part of the episode.
For shows with a host and guests, speaker labels make the transcript readable. Speaker identification is available on paid plans (Plus, Unlimited) or with a Minute Pack, not on the Free plan or the no-sign-up trial. You can set the number of speakers if you know it, then rename Speaker 1 to the host's name once and the change applies across the episode. Hosts with similar voices, or a lot of crosstalk, will need a quick check.
To draft show notes, use the AI prompt export. It copies the transcript with a prompt you can paste into ChatGPT or Claude to write a summary, title ideas or a list of topics. RocketScribe does not generate chapters or summaries itself. Publishing a full transcript on your episode page also makes the episode readable for people who are deaf or hard of hearing, and gives search engines text to index.
Exports for each place you publish
Each export lets you switch timestamps and speaker names on or off, so one transcript can serve several uses.
- SRT or VTT: captions for a video version on YouTube or social platforms. Upload the file as a caption track; RocketScribe does not burn subtitles into video.
- TXT or the Copy button: paste into your website, show notes or newsletter.
- DOCX: an editable transcript for an editor or for turning the episode into a blog post. Turn timestamps off and speaker names on for a clean read.
- Read-only share link: send the transcript to a guest to check quotes before release.
Back catalogues and long episodes
Free accounts have a per-file length cap, which may be shorter than a long-form episode. Paid plans allow much longer files and batch uploads, which helps when you want transcripts for a back catalogue. Use a folder per show or season and bulk export when you are done. Shows in other languages work too, with more than 90 languages supported and mixed Chinese and English handled.