How to transcribe and edit a podcast with AI
Short answer
Record in Riverside, clean the audio with Adobe Podcast or Descript Studio Sound, edit by deleting words from the transcript in Descript, level the loudness with Auphonic, and let AI write show notes and cut social clips. A solo podcaster can do all of it for about $16 to $35 a month, and free tiers cover a first episode.
AI has turned podcast editing from hours of waveform trimming into something closer to editing a document. Here's a complete workflow, with the tool and price for each step, based on each tool's pricing page (checked on the date above).
1. Record each speaker on a separate track
Riverside records remote guests locally on their own computers, so each person gets a clean, separate track even if the internet connection drops. Pro ($24/mo yearly, $29 monthly) includes 5 hours a month of separate tracks, transcription and text-based editing. The free plan includes 2 hours of recording in total, with a watermark.
2. Clean up the audio
- Adobe Podcast Enhance Speech: upload a file and it removes background noise and echo so voices sound studio-recorded. Free for files up to 30 minutes; Premium (reportedly $9.99/mo) handles longer files.
- Descript Studio Sound and Riverside Magic Audio do the same inside those apps.
3. Transcribe and edit by text
In Descript, your recording becomes a transcript; deleting a sentence from the text deletes it from the audio and video. One click removes "um", "uh" and long pauses. Plans start with a free tier (60 minutes a month, watermarked exports) and Hobbyist at about $16/mo billed yearly (10 hours a month, Studio Sound, no watermark).
Private option: MacWhisper transcribes on your own Mac using OpenAI's open-source Whisper model. The free version uses smaller models; Pro is a one-time purchase.
4. Level the loudness
Auphonic automatically balances levels between speakers, reduces noise and hits a loudness target such as -16 LUFS. The free plan processes 2 hours a month.
5. Show notes, chapters and clips
Riverside and Descript both generate show notes, chapter markers and short social clips. You can also paste the transcript into a chatbot and ask for show notes; see how to write better prompts.
6. Captions
Export an SRT caption file from Descript, Riverside or MacWhisper and upload it with your video to YouTube.
What it costs
| Setup | Tools | Monthly cost |
|---|---|---|
| Free first episode | Riverside free, Adobe Enhance free, Descript free, Auphonic free | $0 |
| Solo podcaster | Descript Hobbyist + free Adobe Enhance + free Auphonic | ~$16 to $24 |
| Interview show | Riverside Pro (record, edit, clips) + free Auphonic | ~$24 to $29 |
Need an AI voice for intros or ads? See ElevenLabs vs Murf vs Speechify.
Frequently asked questions
What is the best AI tool for editing a podcast?
Descript is the most popular, because you edit audio and video by editing the transcript like a document, and it removes filler words and cleans audio automatically. Riverside offers similar text-based editing built into its recording platform.
Can I transcribe a podcast for free?
Yes. Descript's free plan includes 60 minutes of media a month, and on a Mac the free version of MacWhisper transcribes on your own computer using OpenAI's open-source Whisper model.
How loud should a podcast be?
A common target is about -16 LUFS for stereo podcasts. Auphonic can level your episode to a target like that automatically.