AI Credits in Podcast Software: What They Are and How They Work

If you’ve opened your podcast software’s billing page lately and found words like “AI credits,” “media minutes,” or “production credits” where a plain monthly price used to be, you’re seeing a shift that’s moved through the whole category in about a year. Most of these tools didn’t raise their headline price. They did something quieter: they changed how the meter runs. Here’s what AI credits actually are, why they make your bill harder to predict, and what to do if you’d rather not think about it every time you edit.

What AI Credits Actually Are

A traditional subscription is simple. You pay a set price, and within your plan’s limits, you use the tool as much as you want. The amount charged to you is the same every month, and you can budget around it.

A credit system adds a meter on top of that. You still pay a set monthly fee, and it still buys you access to the tool, but now the plan comes with an allowance of credits, and certain features spend those credits each time you use them. Plenty of things don’t touch credits at all, so this isn’t a case of paying only for credits. It’s the specific features, usually the AI ones, that draw them down: run an AI editing pass, generate a few clips, enhance the sound, and the allowance empties. Different actions cost different amounts, and the tools are often vague about the exact price of each, so you don’t really know what a given task will cost until you watch the meter drop. When the credits are gone, those AI features pause until you buy more or your next cycle resets, while the rest of the plan keeps working.

Several tools layer a second meter on top called media minutes or media hours. This one counts raw input: every minute of audio or video you upload or record into the tool, regardless of whether you transcribe it, and regardless of whether it survives into the final cut. Record a loose two-hour session to keep twenty minutes of it, and the meter still charges you for the full two hours you brought in.

Put the two together and you have a subscription whose real cost depends on how you work that month, not on the dollar amount posted next to your plan.

Which Podcast Tools Use Them Now

Some of the best-known names have moved onto credits, each a little differently.

Descript made the biggest change. In its September 2025 overhaul, it replaced the old transcription-hour allowances with two meters: media minutes, which count everything you upload or record, and AI credits, which are spent by AI actions like its Underlord assistant. Unused amounts don’t roll over month to month, and when you run out, you buy top-ups that themselves expire after a year. The Studio Sound feature used to be included in the subscription. Now, it spends AI credits every time you run it, on top of the monthly fee you already pay. Replying to a longtime user, a Descript staff member noted that traditional text editing still “uses zero AI credits,” while features “like Studio Sound and remove filler words are unlikely to break the AI credit bank.” Descript plans do come with some credits, so it isn’t a separate charge from day one, but it shows some features that were originally part of the subscription now draw from the AI credit balance.

Riverside meters a narrower set of features. By its own account, AI credits power its premium AI tools. These include things like AI translation, AI B-roll, its AI Twin avatars, and voice cloning, while ordinary recording and editing don’t touch them. The part to watch is in the fine print: Riverside says these credits are purchased separately from your subscription and don’t renew monthly, so the moment you rely on the AI features, you’re topping up a balance on the side. And it drains fast: one minute of AI translation costs 100 credits, and five seconds of AI B-roll costs that same 100.

Zencastr technically runs on credits too, it’s just hiding the meter for now. Its free Vibecastr tier advertises “unlimited production credits,” but its own announcement adds four quiet words: “for a limited time.” The credit system is built in and switched off temporarily, which is a different thing from not having one.

Adobe Podcast belongs on the list too, now that it runs on the same currency. Its $19.99-a-month plan includes 4,000 generative credits, the same currency behind Adobe’s creative AI across Firefly and Creative Cloud. To get more, you’re routed to the Firefly Pro plan with Adobe Podcast Premium. Adobe is blunt about how fast those credits go: its add-on pricing sells 2,000 credits for $9.99 a month and pegs that at roughly six minutes of audio or video translation, so a 4,000-credit monthly allowance is a thinner cushion than it sounds once you lean on the AI tools.

The pattern is what matters. Across tools that used to sell a flat monthly seat, the AI features now sit behind a meter.

The Catch: Why Your Bill Gets Unpredictable

The trouble isn’t that credits exist. It’s that most people can’t tell how fast they’ll spend them, and the tools don’t make it easy to find out until you’re already spending.

The reactions bear this out. On the Descript subreddit, one user upgraded from a legacy plan hoping for something better and instead said that “some features are eating my credits so much”, with AI credits consumed far faster than expected. A commenter on the same thread put the shock plainly: “I blew through a month of credits in a few hours today on the creator plan, working on a 1-minute video.” Another said the AI assistant ate credits so quickly on a batch of conference recordings that it was “costing me a pretty penny,” and resolved to do as much as possible by hand before touching the AI.

The confusion compounds it. A podcaster on a legacy Descript plan described the rollout messaging as a miss: a notice that a price increase was coming without saying what the new price would be, on top of uncertainty about the difference between media minutes, AI credits, and transcription, and whether their plan would still cover a workflow that used to be straightforward. When you can’t tell what you’re buying or what it will cost, budgeting stops being possible.

That’s the core issue. A flat plan answers “what will this cost me” before you start. A credit meter answers it only after you’ve spent, and no rollover means anything you didn’t use is gone.

And when the meter looks switched off, it’s usually just postponed. Zencastr is the clearest example: its production credits are “unlimited” today, but only “for a limited time.” Build your workflow around that free allowance, and you’re relying on a meter that simply hasn’t started charging yet.

Is This Actually a Problem?

AI credits aren’t a scam and metered pricing isn’t automatically bad.

If you lean on AI editing and it saves you hours you’d otherwise pay for, a credit system can be reasonable. For very light users, it can even come out cheaper than a flat plan. Charging by usage is a normal way to price a feature that costs real money to run every time it’s used. None of that is offensive on its own.

The problems are narrower, and there are three. The first is predictability, which we’ve covered: you can’t easily plan around a bill that moves with your workload. The second is that the meter often sits on features you may not want in the first place. If your plan is to record clean audio and edit it yourself, you’re now paying into a system built around AI tools you’ll barely touch, and whether that AI is worth trusting at all is its own question. The third is that this is landing on existing customers, who signed up for one deal and are being moved onto another, which is why the reaction has been sharper than a new-tool launch would get. It isn’t only about credits, either. Riverside users describe the same shape of change from a different angle, features they relied on moved behind a much pricier Business tier, with one longtime subscriber forced to delete studios their Pro plan used to include and calling the jump “from $400/year to $4800” reason enough to leave for good.

So it’s not that credits are evil. It’s that they quietly turn a predictable tool into an unpredictable one, and they assume you want to be paying for AI at all.

How to Keep Your Costs Predictable

If you like the AI features and want to stay, the practical move is to learn which actions draw credits and do the cheap-to-do ones by hand. Cutting, trimming, and basic edits rarely need an AI pass. Save the credits for the things only the AI does, watch the media-minute meter by not dumping hours of raw footage you won’t use, and check whether top-ups or a higher tier is actually cheaper than the overage.

If you’d rather not manage a meter at all, the cleaner answer is to step outside the model. Record with a tool that charges a flat price and hands you clean, separate tracks, then edit wherever you like. That’s the approach SureTake is built around: one predictable monthly price and no AI credits. You record, you download each person’s track, and the cost of your plan and its allowances are clearly outlined. If you want the AI tools, plenty of editors offer them separately, and you can pay for those on their own terms instead of bundling them into your recorder. For people leaving a metered all-in-one, our best remote podcast recording software guide compares the main options depending on how you record and edit.

Frequently Asked Questions

The Bottom Line

AI credits are how the podcasting tools priced their AI push, and once you see the meter it’s hard to unsee. For some people the trade is fine. For a lot of others, it turned a tool they could budget for into one they can’t, and stacked a second charge on the raw footage they bring in. If that’s the part that bothers you, the fix isn’t a cleverer way to ration credits, it’s choosing tools that don’t meter you in the first place.