Why we’re moving from credits to talk time.

Credits made sense to the system

Credits are a very normal way to price AI. They map neatly to underlying usage, model cost, words, tokens, or requests. From the product side, the logic can be perfectly rational.

Then you watch a user try to answer a simple question: how much can I actually use this?

Suddenly they are converting credits into words, words into minutes, Basic into Pro, and a technical accounting unit has become part of the user experience. The pricing model may be correct and still be doing the wrong job.

People already have a unit for speaking

If the product is voice-first, the most natural unit is time. People understand what ten minutes of talking feels like. They can imagine forty-five minutes across a week. They can compare one hour of talk time with the moments in their own workday.

That does not mean the economics underneath disappear. They still matter, a lot. Different modes can have different costs, multilingual processing can affect infrastructure, and usage still has to remain sustainable. But those are product-health problems we should solve behind the interface wherever possible.

The customer should not need a miniature finance model just to decide whether to speak.

A cool feature is only healthy if people value it, use it, and the economics still make sense at scale.

Clear pricing changes behaviour

Pricing is not only a checkout decision. It changes how confidently people use the product. If every action feels like an invisible meter is running, people ration experimentation. They save the product for moments that feel ‘important enough.’ That can be reasonable for expensive infrastructure, but it is dangerous for a product trying to become habit.

A clearer talk-time model lets people understand the trade-off before they begin. They can see what they have, what they are likely to use, and when a top-up makes sense. That transparency is good UX, but it is also good business: confusion creates support cost, hesitation and mistrust.

Product health is where three stories agree

I think about product health as three stories that need to agree: what the user values, how they actually use the product, and what it costs to serve that usage.

If the user loves something that becomes impossible to sustain economically, we have a problem. If the economics look great but people avoid the feature, we also have a problem. If the price is technically fair but impossible to understand, the product still has a problem.

The best pricing system is not the cleverest one. It is the one that makes value legible while keeping the business healthy enough to keep improving the product.

Simple outside, rigorous underneath

This is the same principle we use elsewhere in Lungoor: complexity belongs behind the product. A voice command can feel simple because the hard work happens underneath. Pricing should follow the same rule.

So the direction is straightforward: let people think in talk time, show them usage in a unit they recognise, and keep the infrastructure math where it belongs, with us.