Pricing an Indie App

Pricing is where most indie founders lose their nerve. We pour months into a product, then price it like we are apologizing for charging at all. I have done it. The instinct to set it cheap, to add a generous free tier, to discount before anyone even objects, comes from fear, not from any real understanding of value. When you add AI costs into the mix, that fear becomes financially dangerous.

I want to talk about pricing the way we actually think about it at Sépia, where every inference call costs us real money and a mispriced plan does not just leave money on the table, it loses money on every active user.

Cheap is not a strategy, it is a hiding place

Low prices feel safe. If nobody complains about the price, you reason, you must have it right. But silence on price is not validation. It often means you have attracted the most price-sensitive, least loyal users, the ones who will churn the moment something newer and cheaper appears.

The deeper problem is that a low price caps the kind of business you can build. If your app costs a dollar a month and each user costs you forty cents in AI inference, you have built a machine that converts attention into a slightly smaller pile of money. No amount of growth fixes a broken unit economic.

If you are not slightly uncomfortable with your price, it is probably too low.

Price the outcome, not the bytes

The trap is to price based on what the app costs you to run. That is accounting, not pricing. Users do not care about your server bill. They care about what they get. PrettyType is not selling a face analysis, it is selling a moment of clarity about how you present yourself. MyoScore is not selling a number, it is selling an honest answer to a question people are quietly anxious about.

When you price the outcome, the number can be far higher than the cost, and it should be. The gap between cost and price is not greed, it is the room you need to survive, to keep the lights on, to fund the next version.

Free tiers are a cost, design them like one

For a normal app a free tier is mostly a marketing expense. For an AI app it is a direct, ongoing cost, because every free user burns inference. We learned to design our free experience the way you would design a sample, not a buffet.

• Make the free tier complete, not generous. Let a user feel the full magic once or twice, not unlimited times.
• Cap on the expensive action. If inference is the cost, the limit belongs on inference, not on some cosmetic feature nobody cares about.
• Use scarcity to communicate value. A free quota that resets every few days tells the user this thing is worth something, which makes the paid tier an easy yes.

On PrettyType we settled on a tight free quota precisely because the analysis costs real money and feels valuable. A limited free taste converts better than an unlimited free crutch, and it protects the margin at the same time.

Charge before you are ready

The most common pricing mistake is waiting. Founders tell themselves they will add payments once the product is perfect, once they have more users, once they feel they deserve it. That day never comes, and meanwhile they have trained an entire user base to expect everything for free.

Put a price on it early, even if it is wrong. A wrong price you can learn from is infinitely more useful than a free product that teaches you nothing about willingness to pay. Pricing is a conversation with the market, and you cannot have the conversation if you never open your mouth.

Raise prices, it is allowed

Here is the thing nobody tells indie founders: you are allowed to raise your price. Existing users can be grandfathered. New users will pay the new number and never know the old one existed. We have raised prices and watched conversion stay flat or improve, because a higher price signals a more serious product.

Pricing is not a one-time decision you make in a panic before launch. It is a living dial you adjust as you learn what your work is worth. Stop apologizing for charging. If the product is good, the price is a feature, not a barrier.