Will Plainly raise its price?

$5 is the model, not an introductory offer. An everyday question costs a fraction of the subscription to answer on an efficient model, and sending each request to the model it needs keeps it that way. I do not plan to raise it.

Could Plainly shut down?

A cheap AI app usually dies when its funding stops. Plainly does not run on funding. The subscription covers what the service costs to run, so it stands on its own revenue.

Why does Plainly not run ads?

Ads only work on a free product, where attention is the inventory and the answers can be shaped around a sponsor. Plainly is paid, so there is nothing to sell and no reason to bend an answer.

Does Plainly sell my chats or train on them?

No. Your chats stay private and are not used to train AI models. A paid product does not have to trade your data for the service.

How do AI apps make money? The $5 answer

Q: How do AI apps make money?

Nearly all of them use one of five methods: a monthly subscription, advertising, your data used to train models, API markups, or credit metering. Plainly uses the first only: $5 a month for unlimited chat, with no ads and no training on your chats.

AI apps earn from subscriptions, ads, training data, API markups or credit metering. Plainly earns with the subscription alone: the arithmetic behind $5.

How do AI apps make money?

Five methods cover almost every AI app: a subscription, advertising, your data used for training, API markups, or credit metering. Plainly uses one, a flat $5 a month, because the other four cost the user something.

The five ways, named plainly

How it earns	How it works	Plainly
Subscription	One flat monthly price for access	Yes, $5 a month, unlimited chat
Advertising	Free access, ads beside the answers; your questions become inventory	No
Training data	Free or cheap access in exchange for conversations that train the provider's models	No
API markups	Charge per token, marked up over what the model costs the app	No
Credit metering	Buy credits, spend them per message, so the bill rises with use	No

None of those is a trick; they're real businesses. Advertising is how free apps stay free, and the free version is the inventory rather than a gift. Training data is the same trade paid in conversations. Credit metering is honest about its arithmetic. The cost of each answer lands on whoever asked. API markups are for developers, and priced accordingly.

Plainly uses the plainest of the five: a subscription, $5 a month, unlimited chat, cancel any time. No ads, no credits, no training on your chats. That's the entire revenue line.

Why $5 is enough to run a business

Every answer costs money to produce. Serving the same answer on a frontier model costs an order of magnitude more than serving it on an efficient one, for the kinds of questions most people ask. That gap is the reason cheap AI exists, and it's why $5 can be a business rather than a promotion.

The price only holds if nothing is wasted, so Plainly sends each request to the model the request actually needs. It keeps up with the frontier releases, knows which model suits which kind of question, and deliberately doesn't over-supply: it won't hand you a heavier model than the question needs. A short factual question gets a fast small model. Hard reasoning gets a heavier one.

That's the whole model, and it's deliberately narrow: ruthlessly optimised serving, much of the engineering done with AI, and a product that does one thing instead of twenty. A cheap ChatGPT alternative has one lever: spend on each answer only what the question needs.

Will the price change?

No. $5 is the model, not a launch offer waiting to expire.

It works because answering an everyday question costs a small fraction of the subscription, and the routing keeps it that way. What is left over is modest. Plainly won't make millions, and I never expected it to. It's a good service that still makes money on top of what it costs to run.

That answers the other half of the question: whether it survives. Cheap AI apps usually die when their funding stops. A cheap product funded by investment loses its price the day the funding stops. This one covers its own costs. The risk I'd watch is the market rather than my arithmetic: if efficient models stopped improving or stopped being cheap, the maths would tighten for every cheap app, mine included. Today the trend runs the other way.

What $5 does not buy

The first limit is scope. $5 buys chat, and only chat. No image generation, no agent that works across your desktop, no API, no model picker, no memory across conversations. Those exclusions keep the serving cost down. The catch with cheap AI is a boundary on what you get.

The second is size. This is a small business and it will stay one. It won't outspend the big labs on capability. If your work needs the frontier every day, a $20 plan (ChatGPT Plus, per OpenAI's pricing page, as-of 2026-08-29) is the right call, and a $5 subscription is still the right call for everyone else.

What to do next

If you were asking about the business model, the answer is the table above: one revenue line, a tight serving cost, a price that holds. If you were checking whether $5 is enough for your own use, try it for a month and compare the answers against what you pay for now.