The Quiet Charge
Every subscription service operates on the same assumption: once you're in, you stay in until you actively fight your way out.
Free trials that quietly convert. "Cancel anytime" buttons buried three menus deep. Renewal emails that arrive after the charge, not before. Annual plans that auto-renew at full price while you're on vacation.
The industry has a term for the revenue that comes from users who forgot to cancel: involuntary retention. It sounds clinical. It's actually just a polite way of saying "we charged people who didn't want to pay us."
This isn't a bug in subscription economics. It's the entire business model.
And as of today, we're walking away from it.
What's Changing
Reverie subscriptions no longer auto-renew.
When you purchase a subscription period — whether it's a month or longer — that's exactly what you get. When the period ends, it ends. No silent charges. No surprise renewals. No "we tried to charge your card" emails.
If you want to continue, you choose to continue. Deliberately. Consciously. Because you decided it was worth it.
If you don't renew, nothing happens. Your account stays. Your conversations stay. Your characters stay. You just return to the free tier until you decide otherwise.
This Is Consistent With Everything We Believe
If you've been following our blog, this shouldn't surprise you.
We've told you your conversations belong to you — and built import/export to prove it. We didn't trap your data because we believe users should stay because they want to, not because they can't leave.
We've told you we'd rather under-promise than over-promise — and been transparent about every limitation, every cost reality, every uncertainty.
We've told you technology should enable, not obstruct — and refused to build complexity traps that exploit user confusion.
We've told you we got attacked and it cost us real money — and chose to share the full story instead of a sanitized press release.
Auto-renewal contradicts all of this.
If we believe users should choose platforms based on value rather than lock-in, how can we justify a billing system designed to charge people who stopped choosing us?
If we believe in radical transparency, how can we rely on the opacity of automatic charges?
If we believe technology should enable rather than obstruct, how can we make "stop paying" harder than "start paying"?
We can't. So we stopped.
The Subscription Industry's Dirty Secret
Let's be honest about what auto-renewal really is.
The subscription economy didn't grow because every service became indispensable. It grew because cancellation friction is profitable.
A significant portion of subscription revenue across the industry comes from users who intended to cancel but didn't get around to it. The entire infrastructure of modern subscriptions — free trials, introductory pricing, annual commitments — is engineered around one insight: getting someone to start paying is easy; most people won't bother to stop.
First month 50% off? That's not generosity. That's a bet that you'll forget to cancel before month two hits at full price.
Free 7-day trial? That's not confidence in the product. That's confidence in human inertia.
Annual plan discount? That's not a reward for commitment. That's 12 months of guaranteed revenue from someone who might have left after month 3.
We don't want revenue from people who forgot about us. That's not a business we're interested in building.
What This Actually Costs Us
We're not going to pretend this is painless.
The metrics will look worse. MRR becomes less predictable. Retention rates drop — at least the way the industry measures them. Investor-friendly charts get less smooth.
Revenue will decrease. Some percentage of users who would have auto-renewed will simply... not. They'll forget to come back, or decide to take a break, or realize they weren't using the platform enough to justify the cost. Under auto-renewal, we'd have kept their money. Now we won't.
We're okay with that.
Because every user who renews manually is a user who actively chose us. Every payment represents a conscious decision: "Yes, this is worth it. I want another month."
That signal is more valuable than any MRR chart. It's honest demand. It's real product-market fit. It's the truth about whether we're building something people actually want.
We'd rather have 1,000 users who choose to pay us every month than 5,000 users who forgot to cancel.
The Metric That Matters
Most subscription businesses optimize for retention — keeping people paying regardless of whether they're getting value.
We're optimizing for re-election.
Every billing cycle, users vote with their wallets: is Reverie still worth it? If yes, they renew. If no, they don't. No friction either way.
This forces us to earn your subscription every single period. We can't coast on inertia. We can't rely on dark patterns. We can't hide behind cancellation flows designed to make you give up.
If our product isn't good enough to make you want to come back, we don't deserve your money.
That's a higher bar. It's also the right bar.
Who This Is For
Not every user will care about this. Some people prefer the convenience of auto-renewal and never think twice about it.
But for everyone who's ever:
- Been charged for a service they forgot they subscribed to
- Spent 20 minutes trying to find a cancel button
- Felt guilty about "wasting" a subscription they weren't using
- Avoided free trials because they knew the conversion trap was coming
This is for you.
Your relationship with Reverie is simple: you pay when you want to use it. You stop paying when you don't. No games, no traps, no "are you sure you want to cancel?" guilt trips.
The Pattern
If you've been paying attention, you'll notice a pattern in how we make decisions:
Data ownership: We could lock your conversations in our platform. We chose to build import/export instead.
Pricing transparency: We could hide our cost structure. We chose to show you exactly how credits work and which models we operate at a loss on.
Security: We could hide our attack story behind a vague "technical difficulties" statement. We chose to publish the full account.
Free access: We could promise unlimited free forever. We chose to be honest that free models might not last.
And now, subscriptions: We could keep auto-renewal and enjoy the predictable revenue. We chose to let you decide every time.
The pattern is simple: when there's a choice between what's better for our metrics and what's more honest to our users, we choose honesty.
Not because we're noble. Because we're building for the long term, and long-term trust compounds faster than short-term revenue.
Try It
Pick a subscription period. Use Reverie. See if it's worth it.
When the period ends, decide again. No pressure. No countdown timers. No "last chance to keep your discount" emails.
Just a simple question: do you want another month?
If yes, welcome back.
If no, see you when you're ready. Your characters will be here waiting — and this time, they might even message you first.
Your subscription. Your choice. Your timeline. That's how it should be.
Update — August 2026: auto-renewal is back
We're leaving the post above exactly as we wrote it in April, because editing it quietly would be worse than the reversal itself.
Card subscriptions auto-renew again. Here's the honest account of why, and what we think we got wrong the first time.
The complaint we didn't expect
We assumed the people who disliked auto-renewal were the ones we'd hear from. Instead, most of the feedback came from users who wanted to keep subscribing and found it genuinely annoying.
Manual renewal meant a lapse. Your subscriber benefits stopped the moment the period ended — usually noticed mid-conversation, on a day you weren't thinking about billing — and getting them back meant going through checkout again. Every month. People who were happy to pay us were being asked to re-prove it on a schedule, and a decent number of them just got tired of it.
That's not the "conscious re-election" we described in April. That's friction pointed at the wrong people.
The part about money
The other reason is straightforward and we'd rather say it than dress it up.
Revenue that resets to zero every month is very hard to plan against. We've written separately about where the money actually goes — the short version is that a lot of our costs are committed and recurring whether or not anyone renews. Committing to those against a revenue line that has to be rebuilt from scratch every thirty days meant either holding large buffers or being cautious in ways that showed up as slower shipping and a more conservative free tier.
Predictable subscription revenue is a meaningful part of what keeps the free tier as generous as it currently is. That's the trade, stated plainly.
What we actually got wrong
Reading the April post back, the error is in the argument, not just the outcome.
We treated auto-renewal and involuntary retention as the same thing. They aren't. Involuntary retention is charging people who don't want to pay — and it's produced by buried cancel buttons, guilt-trip flows, and charges that land before anyone tells you they're coming. Auto-renewal is just a billing mechanism. It becomes predatory when leaving is hard, and it's unremarkable when leaving is easy.
We attacked the mechanism when the thing worth attacking was the friction around it. So that's where the commitment now lives:
- Cancel any time, from billing settings, without talking to anyone.
- Cancelling keeps what you paid for. You keep subscriber access until the end of the period you already bought — cancelling on day two of a month doesn't cost you the other twenty-eight.
- Your renewal date is always visible in billing settings, not buried.
- Nothing is charged after you cancel.
Two places we moved, stated outright
We'd rather name these than let you find them.
The cancellation flow now offers a discount if you're leaving because of price — once per subscription. The April post specifically mocked "last chance to keep your discount" as a dark pattern. You can decline it and cancelling still completes. But if you hold us to April's wording, this is a place we changed our minds, and we're not going to pretend it's the same thing we promised.
We don't email before every renewal. Non-renewing subscriptions get a reminder three days before they lapse; auto-renewing ones show their date in billing settings instead. If enough people want a pre-charge email, we'll build it — that one's a straightforward fix and we're open to being pushed on it.
What hasn't changed
Everything the April post said about data ownership, export, and not trapping people still holds, and none of it depended on the billing mechanism. You can still take your conversations and leave whenever you want.
We were probably too pleased with ourselves in April. The principle was right; the implementation punished the wrong users and made the business harder to run in ways that reached everyone, including people who never subscribed at all. Reversing it is the more honest outcome, and pretending we'd been consistent all along would have been the cheaper story to tell.
If you preferred it the old way, say so — that feedback is why we changed it the first time.