I produce three podcasts. One’s mine, two are clients’. That’s roughly 12 episodes per month, every month, for the past two years.

My workflow has changed dramatically since episode one. What started as a 6-hour edit-per-episode nightmare is now under 2 hours, including show notes and upload.

Here’s exactly what I do.

The Stack at a Glance

Stage	Tool	Cost
Recording	Riverside.fm	$15/mo
Editing	Descript	$24/mo
Audio cleanup	iZotope RX (elements)	$129 one-time
Hosting	Transistor	$19/mo
Show notes	Notion + ChatGPT	Free/$20

Monthly total: $78 + $20 ChatGPT Setup time: One weekend to learn, refined over 20 episodes

Why This Setup

I optimized for one thing: getting episodes out consistently without burning out.

That meant accepting “good enough” audio instead of perfect, automating everything that can be automated, and standardizing the process so I don’t make decisions every episode.

Creative energy goes into content, not production logistics.

Stage 1: Recording

The Tool: Riverside.fm

Riverside records each participant locally in high quality, then uploads. This matters because Zoom audio sounds like Zoom audio—compressed, inconsistent, bad when internet hiccups.

Riverside gives me separate tracks per person, full quality, with video if I want it.

Why not Zencastr or SquadCast? I’ve tried both. Riverside’s upload reliability is better in my experience. Fewer failed recordings, faster sync.

Why not record locally and share files? Works fine for solo or two-person same-room recordings. Falls apart with remote guests who don’t understand audio settings.

The Recording Process

1. Riverside link sent to guest 24 hours before, with instructions to use Chrome
2. 5-minute buffer at call start for tech check
3. Record with video (even for audio-only pods—facial expressions help me edit later)
4. Brief chat after recording to capture anything we missed
5. Riverside processes and delivers files within 30 minutes

Time: 1 hour (interview length varies, but 45-60 minutes is my target)

Stage 2: Editing

The Tool: Descript

I covered Descript in depth in another post, but for podcast editing specifically: the text-based editing changes everything.

Import audio. Transcription happens automatically. Edit the transcript like a document. Delete the tangent? Highlight, delete. Move a section? Cut and paste.

The Editing Process

1. Import both audio tracks to Descript
2. Let transcription complete (5-10 minutes)
3. Read through transcript, deleting:
   • Ums, uhs, filler (automated “filler word” removal gets most of this)
   • False starts and repeated sentences
   • Tangents that don’t serve the episode
   • Any sections the guest asked to remove
4. Timeline cleanup:
   • Tighten gaps between speakers
   • Fix any cross-talk by choosing who to hear
   • Adjust levels if one track is significantly louder
5. Add markers for show notes timestamps

Time: 45-60 minutes for a 45-minute episode final length

What I Don’t Do

I don’t obsess over perfect edits. A slightly awkward transition is fine. A small mouth noise that slipped through is fine. Listeners don’t notice what I notice.

The difference between a 2-hour edit and a 6-hour edit is mostly perfectionism, and my listeners can’t hear the difference.

Stage 3: Audio Cleanup

The Tool: iZotope RX Elements

Before exporting from Descript, I run the audio through iZotope for:

• Noise reduction (air conditioning hum, room tone)
• De-clicking (mouth noises)
• Loudness normalization

Why not Descript’s built-in Studio Sound? It works okay. iZotope works better, especially on problematic audio. The one-time $129 cost paid for itself in reduced headaches.

The Process

1. Export from Descript as separate tracks
2. Open in Audacity (free) with iZotope as plugin
3. Run noise reduction preset on each track
4. Run de-click if needed
5. Normalize loudness to -16 LUFS (podcast standard)
6. Export final mix

Time: 15-20 minutes

Stage 4: Show Notes and Upload

The Tool: Notion + ChatGPT

I used to write show notes from scratch. Now I paste the transcript into ChatGPT with this prompt:

“Create podcast show notes from this transcript. Include: 3-sentence summary, key topics discussed as bullet points, timestamps for major sections, 5 relevant keywords.”

It gets me 80% there. I clean up the output, add links, and adjust the voice.

Show notes live in Notion, linked to the episode entry in my content calendar.

Upload to Transistor

1. Upload MP3 to Transistor
2. Paste title, description, show notes
3. Set publish date
4. Add chapter markers (Transistor supports these)
5. Schedule social posts in Transistor (optional)

Time: 20-30 minutes

Total Time Per Episode

Stage	Time
Recording	60 min
Editing	45-60 min
Audio cleanup	15-20 min
Show notes + upload	20-30 min
Total	2.5-3 hours

For a client episode where I’m not on the call, subtract the recording time—I just receive files and process.

Where Things Break Down

Bad Remote Audio

When a guest records in a bathroom-echo room with AirPods, no amount of post-processing fixes it. iZotope helps. It doesn’t perform miracles.

I’ve started sending a “recording checklist” to guests: quiet room, wired headphones if possible, phone on silent. Cuts audio issues by half.

Episode Backlogs

If I skip a week, episodes pile up fast. Three podcasts means falling behind creates a 6+ episode backlog quickly.

My fix: batch recording days. Two episodes per podcast, recorded same week, edited same week. Creates buffer without killing momentum.

Guest No-Shows

Happens more than I’d like. I keep a list of “backup” solo episode topics that I can record on short notice.

Budget Alternatives

The full stack costs ~$100/month. Here’s the same workflow cheaper:

Stage	Tool	Cost
Recording	Zencastr free tier	$0
Editing	Audacity	$0
Audio cleanup	Audacity built-in	$0
Hosting	Anchor	$0
Show notes	Manual	$0
Total		$0

This works. Editing takes 3x longer without Descript’s transcription. Audacity’s learning curve is steeper. Anchor has limitations on analytics and monetization.

But if $100/month isn’t in the budget, you can absolutely produce a quality podcast with free tools.

What I’d Change

If money weren’t a factor:

Upgrade Descript to Team plan for better collaboration with clients who want to review edits.

Add a proper mixer for in-person recordings when I have guests in my space.

Hire an editor for the client podcasts. My time is worth more than the edit time at this volume.

For now, the current setup balances cost, quality, and speed. I revisit it every six months to see if anything’s changed.

---

This is my actual workflow as of October 2024. 150+ episodes produced across three shows.