Descript vs Riverside: they do different halves of the same job
Most comparisons of these two rank them. Ranking them is the mistake — one of them records the conversation and the other one assembles it, so the useful question is which half you cannot do yet.
We make neither of these, and there is no affiliate link on this page
InTheClips is a clipping tool: you give it a finished recording and it tells you which passages are worth posting. That makes us a neighbour of both products on this page and a competitor to neither on the thing they are being compared for — Riverside records, Descript edits, and we do not do either.
It also means we earn nothing whichever one you pick. Neither company is an affiliate partner, no link below is tracked, and the recommendation does not change based on who pays. We appear once, in one section near the bottom, for the job that is genuinely ours, and that section is labelled.
Choose by which half of the job you cannot do yet
If the recording does not exist yet, that is Riverside’s half. It is a studio that runs in a browser: your guest joins from a link, and each side of the conversation is captured on their own machine rather than through the call. That is the reason it exists — a video call compresses and drops what it needs to in order to stay live, and no editor can put back what was never recorded. If your show is remote interviews, this is the half that has to be solved first.
If the recording exists and has to be turned into an episode, that is Descript’s half. It is an editor whose timeline is the transcript: you delete a sentence and the video under it goes with it. For anyone who edits by reading rather than by scrubbing, that is a genuinely different working speed, and it is the reason people describe it as the first video editor they did not have to learn.
Most shows end up wanting both, and that is not a hedge. They sit in a sequence rather than in a contest — capture, then assemble — and each one is weakest at the other’s half. The question worth asking is not which of the two is better but which one you are currently doing by hand, badly, on a Sunday evening.
What each one is built around
Sentences rather than ticks, and the same weight on both columns. A tick against a product name is a claim about software we do not run, and a table that leans is one you would be right to stop reading.
| What differs | Descript | Riverside |
|---|---|---|
| The problem it was built for | Turning a recording into a finished piece. The transcript is the document, and editing the words edits the video. | Getting a good recording out of a remote conversation in the first place. |
| Where the audio and video come from | Files you already have, plus its own screen and microphone capture for the sort of recording you make alone. | The session itself. Each participant is captured on their own machine and the files are uploaded afterwards, so the quality does not depend on the call holding up. |
| Bringing in a guest | Not what it is for. A guest recording arrives as a file, from wherever it was made. | A link. That is the product: the guest opens it in a browser, and their side is recorded at their end. |
| How you make changes | By editing text. Removing a paragraph removes the seconds it occupied, and the machinery of a normal editor is underneath if you want it. | In its own editor, which is built to get the recording it just made into shape rather than to be the place you cut everything. |
| What it hands you at the end | An exported episode, in the shape you assembled. | The tracks, and an edit of them. Many shows take those files somewhere else afterwards — including into Descript. |
| Short clips out of a long recording | Both of them offer this, and neither of them is only this. | Both of them offer this, and neither of them is only this. |
| What it costs | Its own tiers, on its own site. A price copied here would be stale by the time you read it. | The same. Read it from the source, on the day you decide. |
The last two rows are deliberately unhelpful as a tiebreak, and they are the two most comparison pages lean hardest on. Both columns are from each product’s own public description of itself, read on . Neither company had any part in this page.
Three situations, and what each one points at
Written as the answer somebody with nothing to sell would give across a table.
Your show is remote interviews
Solve the capture first. An episode recorded through a call is a problem no editor can undo, and every hour spent choosing an editor before that is an hour spent on the wrong half. Riverside, then whatever you like for the edit.
You record alone, or in a room
The capture is already solved by the camera and the microphone in front of you, and the work is all downstream. That points at the editor, and at whichever one you can move fastest in — which, for people who would rather read than scrub, tends to be Descript.
The episodes exist and nothing gets posted
Then neither of these is the bottleneck. The unfinished job is finding the forty seconds inside two hours that is worth posting, and that is a reading problem rather than a recording or an editing one.
One section, at the bottom, on the job neither of the two above is bought for. If the first two cases are yours, this one is not, and the page above is the whole answer.
InTheClips takes a finished video — recorded anywhere, edited anywhere — and reads it end to end. There is no studio in it and no timeline editor: it does not compete with either half above, it runs after them.
Every passage it would cut arrives as a proposal: four scores, the timecodes, and the sentences that made it a candidate. You keep or you pass, and nothing is rendered until you do.
Nobody can, and a page that claims it should be read sceptically. What is checkable is that you see the list and the reasoning before anything is made, which is the part you can judge for yourself in one episode.
The link box takes youtube.com and youtu.be and refuses the rest, so a recording sitting in Drive, Zoom or Dropbox is an export and an upload first. Imports are capped by the day — three on the trial, a hundred at the top — and nothing is posted anywhere on your behalf.
Questions
The questions this search turns into once somebody realises the two are not substitutes.
Can I use Descript and Riverside together?
Yes, and it is the common arrangement: record the session in one, take the resulting files into the other to assemble the episode. They are consecutive rather than competing, which is why a straight ranking of the two rarely helps anyone.
Does Descript record remote guests?
It records what is in front of you — your screen and your microphone — which covers a solo piece or a walkthrough. Capturing a second person at their end, on their own machine, is the specific problem Riverside is built around, and it is the clearest line between the two.
Which one is better for a podcast?
It depends on which half of the podcast is currently hard. If getting a clean recording out of a remote guest is the part that goes wrong, the recorder is the answer. If the recordings are fine and the edit is what eats the week, the editor is. A show where both are already fine does not need either of them for the clips.
Which one is cheaper?
Both publish their tiers on their own sites, and any number written here would be out of date on some Tuesday without this page noticing. Read them from the source on the day you decide — and compare what a tier includes, because the units they are sold in are not the same unit.
Do either of them make the short vertical clips?
Both offer a feature that does, and neither product exists mainly for it. If cutting shorts out of long recordings is the whole of the job you are trying to solve, it is worth comparing tools built for that specifically rather than choosing between two tools that also do it.
I already have both. What is still missing?
Usually the reading. A two-hour episode still has to be scrubbed by somebody to find the four passages worth posting, and that somebody is you — which is the job we built ours for, and the only reason we have a section on this page at all.
Other comparisons written this way
Same rules: sentences rather than ticks, a date on the table, and no affiliate link on any of them.
