Riverside fixes the problem that quietly ruins most remote podcasts: when you record a conversation over a video call, you are capturing compressed internet audio, not how anyone's voice actually sounds. If you record with remote guests, this is the tool that makes everything downstream easier. The whole platform is organized around one technical idea, recording each person locally instead of over the wire, and once you hear the difference it is hard to go back to recording a call.
What it does best
Local recording. Riverside captures each person's audio and video on their own device, then uploads the clean, uncompressed files afterward, so a guest's quality is exactly what their microphone heard, no matter how shaky their connection was during the call. This matters because a normal video call does two destructive things at once: it compresses the audio to fit the bandwidth, and it drops quality further whenever the connection wobbles. You hear the result as that thin, slightly underwater sound that marks an amateur remote episode. Recording locally sidesteps both. The microphone's clean signal is saved on the guest's machine and synced up later, so a momentary drop in their Wi-Fi shows up as a glitch in the live monitor, not in the final file.
The difference is audible. A guest recorded through Riverside in a home office often sounds better than the same guest on a normal Zoom call with a nice mic, because you are not stacking internet compression on top of the microphone. That is the single best reason to use it, and it is the reason it makes everything after the recording easier: you start your edit from clean source files instead of trying to rescue compressed ones. The AI transcription runs automatically afterward and is accurate enough that most people use it as their editing reference, scrubbing to the right moment by reading rather than scrubbing the waveform blind.
Pricing and what you actually get
The free plan is a real trial. It caps recordings at two hours and stamps a watermark on video exports, which is enough to run a session end to end and hear the local-recording difference for yourself before paying. The Standard plan around $19/month lifts that to roughly five hours of recording and removes the watermark, so you can ship real episodes from it. The one thing to do before committing is simple arithmetic: count how many hours you actually record in a month, including the takes you throw away, and confirm five hours covers it. A weekly show with long sessions and re-records can quietly outgrow that ceiling, and you want to know that going in rather than mid-month.
Where it falls short
Magic Clips, the feature that auto-finds highlight moments and cuts short-form clips, is handy but uneven. It does pick genuinely engaging moments, and it also regularly misses context-dependent humor or the payoff that only lands because of a long setup it cut away, so treat its picks as a starting shortlist rather than finished clips. The AI is good at spotting energy and bad at spotting why a moment is funny, which is exactly the kind of judgment that makes a clip worth posting.
The other thing to know is that quality depends on your guest's setup. Local recording only works if their browser permissions are granted and their device has storage room for the file, and a guest who blocks the mic permission or has a full drive can leave you with a degraded fallback recording instead of the clean local one. The platform is doing the right thing technically, but it is leaning on each guest's machine to cooperate, and not every guest is technical.
How it compares
Against recording a plain Zoom or Google Meet call, Riverside wins on the thing that matters most, source quality, because those tools capture the compressed stream and Riverside captures the clean local file. Against a pure audio tool, the draw is that Riverside does synced video and audio together with transcription and clip features attached, so a video-first show does not need a separate stack. Riverside's value rises with distance: the more remote and less technical your guests, the more it earns its keep.
Who it's for
Remote-first podcasters and video interviewers who care about how their guests sound and do not want to coach everyone through pro audio gear. It fits the host who books guests from anywhere and needs each one to sound consistent without shipping them a microphone. If you only ever record in one room in person, a simpler local setup will do the job without the subscription, and you will not be paying for the remote-recording machinery that is the whole point here.
Getting the most out of it
Send guests the link at least 30 minutes early so they can run a quick test, since the local recording only works when their permissions are granted and their storage has room, and the test is where you catch a blocked mic before it costs you an episode. During the call, hit "Mark moment" whenever something clip-worthy happens. Those live timestamps save a lot of hunting in post, and they make Magic Clips less of a gamble, because you already know where the good parts are and can check the AI's picks against your own marks instead of trusting it blind.
