How to Record, Transcribe, and Clean Up a Podcast Interview with Riverside FM
A step-by-step guide to recording studio-quality podcast interviews remotely with Riverside FM , covering setup, local recording, AI transcription, and audio cleanup with Magic Audio.
Recording a podcast interview remotely used to mean making peace with compressed, choppy audio, the kind where one person’s voice drops out whenever their internet hiccups. Riverside FM solved that problem by recording each participant’s audio locally on their own device and uploading it to the cloud afterward. What you get is each person’s track at full quality, regardless of what their connection did mid-interview.
This guide covers the full workflow: setting up a Riverside studio, preparing your guest, running the recording session, using AI transcription, and cleaning up audio with Magic Audio, all the way through to a polished, ready-to-edit file.
Prerequisites
Before your first recording session, you’ll need:
- A Riverside account. The free plan includes two hours of multi-track recording, which is enough to test the workflow. Paid plans start at $24/month (Standard).
- A decent microphone. A USB condenser mic is enough to produce noticeably better audio than a laptop mic. Riverside’s Magic Audio can improve room audio after recording, but it works better when the source recording is already reasonably clean.
- Closed-back headphones. These prevent your guest’s audio from bleeding into your microphone when you’re both on the call.
- A quiet recording space. Background noise, fan hum, air conditioning, street sound, is the main thing that makes home-recorded interviews sound amateur. Soft furnishings (rugs, curtains, a closet full of clothes) reduce echo better than any processing tool.
- A stable internet connection, at least for upload purposes. Because Riverside records locally, a dropped connection won’t destroy the recording quality, but participants need enough bandwidth to finish uploading their tracks after the session ends.
Your guest doesn’t need a Riverside account. They join via a link in their browser.
Step 1: Set Up Your Riverside Studio
A Studio in Riverside is a persistent room you create for a specific show or recurring interview format. You can create as many as your plan allows, and each one stores your past recordings in one place.
- Log in to your Riverside dashboard and click Create a Studio.
- Give it a name (your show name works) and choose whether you’re recording audio only or video + audio. If you want a video podcast or plan to publish to YouTube, choose video.
- Under Studio Settings, set your language. Riverside supports 100+ languages and uses this for transcription accuracy.
- Optionally, set up Studio Branding, logo, background color, and lower thirds if you want the recording interface to look polished on video.
That’s your studio created. You can leave it and come back to it for every episode.
Step 2: Check Your Audio Setup Before the Guest Joins
A quick technical check before your guest joins saves a lot of post-production headaches. Riverside has a built-in mic test at riverside.com/mic-test, use it to confirm your microphone is being picked up correctly.
Things to verify before the session:
- Microphone selected correctly in your browser permissions (not the laptop’s built-in mic)
- Headphones plugged in
- Room is as quiet as it’s going to be
- Camera on and framed reasonably (for video recordings)
One practical note on browser permissions: Riverside runs entirely in the browser, so you’ll need to allow microphone and camera access the first time you open a studio. If permissions are blocked, Riverside shows a prompt to fix it, but it’s faster to resolve this before your guest is waiting.
Step 3: Invite Your Guest
From inside your Studio, click Invite Guest. Riverside generates a unique link. Send it to your guest via email, Slack, DM, or whatever you normally use to communicate.
When your guest clicks the link, they land on a pre-session check page where Riverside walks them through microphone and camera setup in their browser. No download or account is required. The whole join process takes about a minute, and Riverside’s built-in troubleshooting prompts handle the most common issues (wrong mic selected, permissions blocked) without you having to debug anything.
A few things worth telling your guest before they join:
- Use headphones. Audio bleed from laptop speakers into a microphone creates a mess that’s hard to clean up in post.
- Find a quiet room and close the door. This is the single highest-impact thing they can do.
- Use Chrome or Firefox. Safari has known compatibility limitations with recording tools.
- Keep their browser tab active during the recording. Background tabs on mobile devices, in particular, can interrupt the local recording process.
Step 4: Run the Recording Session
Once both of you are in the studio, you’ll see the multi-participant interface. Check that you can both see and hear each other, then click the red Record button to start.
A few things to know about how Riverside records:
Local recording. Riverside saves the audio (and video) to each participant’s device as the session progresses, then progressively uploads to the cloud. This means if either of your internet connections drops during the interview, the recording continues locally and uploads when connectivity resumes. The quality of the final file is not affected by the connection quality, only the upload timing is.
Per-participant tracks. Each person gets their own separate audio (and video) track. When you download the raw files, you’ll have one WAV file per participant rather than a single mixed stereo track. This is the core advantage over recording a Zoom call: you can process each voice independently in your editor, fix volume differences, cut one person’s track without touching the other, and send individual tracks to different processing steps.
48kHz WAV audio. Riverside captures audio at 48kHz WAV by default, lossless, uncompressed, and significantly higher quality than the MP3 or Opus compression that Zoom and Teams apply to call audio.
The Teleprompter. If you have prepared questions or a rough script, Riverside has a built-in teleprompter you can load before recording. It scrolls during the session so you’re not glancing away from the camera to check notes.
During the interview itself: keep an eye on the upload progress indicator in the corner of the interface. Riverside shows a percentage for each participant’s upload. If someone’s upload is stalling (usually a slow connection), the recording itself is still happening locally, it just means their track will finish uploading a few minutes after the session ends.
When you’re done, click Stop Recording. Give everyone a minute to let their tracks finish uploading before closing the browser.
Step 5: Download Your Raw Tracks
After the session ends, your recordings appear in the Studio’s recording list. Click into the recording to access the tracks.
From the Tracks section, you can download:
- Individual audio tracks (WAV) per participant
- Individual video tracks (MP4) per participant, if you recorded video
- A mixed audio file (all participants combined)
For most podcast interviews, you’ll want the individual WAV files. These are what you’ll bring into your editor for cleaning, mixing, and cutting.
If you plan to edit in Riverside’s own text-based editor (covered in the next step), you can also work directly from the recording inside the platform without downloading anything first.
Step 6: Transcribe the Interview
Riverside transcribes recordings automatically in the background once a session ends. By the time you’re ready to review, the transcript is usually ready.
To access it:
- Open the recording in your Studio.
- Click into the editor view, the transcript appears alongside the waveform, with speaker labels per participant.
- Each line of the transcript is clickable, click any word to jump to that point in the audio.
You can also use Riverside’s standalone transcription tool to upload pre-recorded files (not just recordings made in Riverside). The free tool at riverside.com/transcription handles MP3, WAV, MP4, and MOV files, supports 100+ languages, and doesn’t require a login.
Riverside claims 99% accuracy, which holds up reasonably well in practice for standard accents in quiet recording conditions. Accuracy drops with strong accents, heavy technical vocabulary, or noisy recordings. For the latter, run Magic Audio (Step 7) before transcribing an uploaded file, cleaner audio produces better transcription output.
Downloading the transcript:
- TXT format, plain text, good for repurposing into show notes, blog posts, or newsletter copy
- SRT format, timestamped subtitle file, used for adding captions to video exports
Speaker detection works automatically when each participant is on a separate track (which is the default in Riverside sessions). If you upload a single mixed audio file, speaker detection will not separate voices.
Step 7: Clean Up Audio with Magic Audio
Magic Audio is Riverside’s one-click AI audio enhancer. It removes background noise, reduces room echo and reverb, balances volume levels across tracks, and applies voice enhancement, all in a single step.
To apply it:
- Open your recording in Riverside’s editor.
- Click AI Producer in the editing toolbar.
- Scroll to Magic Audio and click Apply.
- Processing takes a few seconds. Play the recording again to hear the difference.
What Magic Audio handles well:
- Fan hum and air conditioning noise
- Mild room echo from untreated spaces
- Volume imbalances between participants (common when one person is closer to their mic)
- Low-level street noise and ambient background sounds
What it doesn’t fix:
- Severe clipping or distortion from speaking too close to the mic, this is an artifact baked into the waveform and no AI tool reverses it cleanly
- Audio bleed from one participant’s track onto another, if someone recorded without headphones
- Very loud or sudden noise events (a door slamming mid-sentence, a phone ringing)
For the best results, apply Magic Audio to each participant’s track separately before exporting, rather than applying it to a mixed file. Separate-track processing gives the AI more to work with per voice.
One limitation to flag: Magic Audio is available on Standard and Pro plans. The free plan includes two hours of multi-track recording but does not include Magic Audio. If you’re evaluating Riverside on the free plan, you can test the transcription workflow but won’t have access to the audio enhancement until you upgrade.
Step 8: Edit and Export
Once transcription and audio cleanup are done, you have two paths:
Edit inside Riverside. The platform has a text-based editor, delete words or paragraphs from the transcript, and the corresponding audio is removed automatically. This is the same paradigm as Descript (covered in the 7 Best AI Tools for Solo Podcasters in 2026 roundup). For podcasters who want a single-platform workflow from recording through to a rough edit, this works well for standard cuts and filler word removal.
Export to an external editor. If you prefer Descript, Audacity, Adobe Audition, or another editor, export the cleaned WAV tracks from Riverside and bring them into your tool of choice. Riverside exports in formats that work with all major audio editors.
For the export:
- In the editor, click Export (or download from the Tracks section).
- Choose your format, WAV for maximum quality, MP3 if file size matters more.
- Download individual tracks or a mixed file depending on what your next tool expects.
If you’re also publishing to video platforms, Riverside’s Pro plan includes direct publishing to Spotify, Apple Podcasts, and YouTube from within the platform, which removes one more step from the distribution chain.
Common Mistakes to Avoid
Not telling guests to wear headphones. This is responsible for more unusable recordings than any technical failure. Audio bleed from speakers into a microphone creates echo that’s genuinely hard to remove cleanly.
Closing the browser before tracks finish uploading. Riverside uploads tracks progressively during the session, but if a participant closes their browser the moment you stop recording, their upload may be incomplete. Wait for the upload indicator to reach 100% before closing.
Skipping the pre-session check. Five minutes of mic and camera testing before the guest joins catches most of the problems (wrong device selected, permissions blocked) that would otherwise interrupt the first 10 minutes of the interview itself.
Recording without headphones on your end. If you’re monitoring audio through speakers during the session, your microphone will pick up your guest’s voice and create an echo on your track.
Applying Magic Audio to a mixed file instead of separate tracks. The enhancement works better per-voice. Export your tracks, process them individually if you’re doing heavy cleanup, then mix afterward.
Quick-Reference: The Riverside Interview Workflow
| Step | What happens | Where |
|---|---|---|
| Studio setup | Create a persistent room for your show | Riverside dashboard |
| Pre-session check | Mic, camera, headphones, quiet room | Before guest joins |
| Guest invite | Share link, no account needed | Studio > Invite Guest |
| Recording | Local recording per participant, lossless WAV | Inside the studio |
| Transcription | Automatic, with speaker labels | Riverside editor |
| Audio cleanup | Magic Audio one-click enhancement | AI Producer toolbar |
| Export | Individual WAV tracks or mixed file | Tracks section |
Pricing Summary
| Plan | Monthly price | Recording | Transcription | Magic Audio |
|---|---|---|---|---|
| Free | $0 | 2 hrs multi-track | Via standalone tool (no speaker labels) | No |
| Standard | $24/month | Unlimited | Unlimited, with speaker labels | Yes |
| Pro | $40/month | Unlimited | Unlimited | Yes + direct publishing |
Prices as of July 2026. Check riverside.com/pricing for current plan details.
Try Riverside FreeFAQ
Do my guests need a Riverside account?
No. Guests join via a browser link with no signup required. Riverside walks them through a quick mic and camera check when they arrive.
What happens if the internet drops mid-recording?
Because Riverside records locally on each participant’s device, a connection drop doesn’t interrupt the recording itself. The track continues saving locally and uploads when connectivity is restored. The quality of the final file is unaffected.
How accurate is Riverside’s transcription?
Riverside claims up to 99% accuracy. In practice, accuracy is high for standard English in reasonable recording conditions. It drops meaningfully with strong accents, technical jargon, or noisy audio. Running Magic Audio before transcribing uploaded files improves results.
Can I use Riverside for solo recordings (no guests)?
Yes. Riverside works for solo recording, you’re simply the only participant. The per-participant track structure is less of an advantage without a guest, so dedicated editing tools like Descript may offer a more complete solo workflow. That tradeoff is covered in detail in the 7 Best AI Tools for Solo Podcasters in 2026 article.
Is Magic Audio available on the free plan?
No. Magic Audio is included on Standard ($24/month) and Pro ($40/month) plans. The free plan covers two hours of multi-track recording and access to the standalone transcription tool, but not audio enhancement.
Can I edit my podcast inside Riverside without a separate editor?
Yes. Riverside has a text-based editor where you delete words or paragraphs from the transcript to cut the corresponding audio. For basic cuts, filler word removal, and a rough edit, it covers most of what you need without leaving the platform.
References
- Riverside recording features and local track capture: riverside.com/recording
- Riverside transcription (99% accuracy, 100+ languages, free standalone tool): riverside.com/transcription
- Magic Audio features and how-to guide: riverside.com/magic-audio
- Riverside pricing (Free, Standard $24/month, Pro $40/month): riverside.com/pricing
- Riverside affiliate program (up to 20% recurring commission): riverside.com/affiliate-program
- Riverside mic test tool: riverside.com/mic-test