06 — Monetisation Strategy

Free, and five packs. Two rails. One rule.

The rule is that safety is never behind money. Every feature is on Free, for everyone, forever, with the Free limits. A pack raises the limits for a year — bigger files, more devices and accounts, larger groups and group calls, unlimited call recordings, cloud backup — and that is the only thing we sell.

The plans

Free, then Basic, Pro, Plus, Max and Ultra. A higher pack includes every pack below it, and a pack lasts a year on every device of the identity that holds it. The price is served from a price book on our own server — the app asks, it does not decide — by price group: India and the lower-middle and low income countries pay the INR group's prices, the rest pay in USD, EUR or GBP. We can change a country's price without shipping a build, and cannot ship two prices in two places.

Nobody has bought a pack yet, so the shares above are ours, not measured. They are weighted to the cheaper packs, because a pack raises limits most people never reach and the cheapest one is the easiest yes. At that mix a paying user spends a year in India and elsewhere, and every projection on this site is built on those two numbers.

Why not cheaper

In trust markets, cheapest is a bad position. A price too low invites exactly the question we want to answer cleanly — then how are you really funded? Threema charges once and is taken seriously. Session is free, donation-funded, and came within days of shutting down in July 2026. Our position is "sustainably priced, and here is the arithmetic". The durability argument is the marketing.

The two rails, and why the difference is most of our margin

A pack bought through an app store pays the store fifteen per cent of the first million dollars a year and thirty per cent after that. The same pack bought through Razorpay — on our website, or in the checkout inside the Android, Mac and Windows apps — pays a payment gateway about three per cent. Same product, same price to the customer, very different economics for us.

RailWhat it costs usWhat the customer does
Razorpay, our own rail2–3%Pays by UPI, card or net banking inside the Android, Mac or Windows app, and the pack switches on by itself; or on our website, and gets a pack code to add in the app. Either way the pack that turns on is blind-signed on their device.
App store15% → 30%Taps buy inside the iPhone app. Simpler, and the only way on an iPhone, so we built it and we pay for it.
On what a paying user spends in India, a store takes a year. Serving a user costs us about a year at a million users. The store is the largest single cost line in this business and it is not infrastructure.

We do not model everyone moving to the cheaper rail. The store's share is each market's iPhone share — in India, in the USD, UK and EEA markets — because an iPhone can only buy through the App Store. At the year-three mix that puts of sales through Razorpay. The calculator on The Pitch lets you set it market by market; after conversion it is the most consequential setting on this site.

Why we can carry free users forever

This is the part that is easy to state and hard for anyone else to copy. Serving a user costs us about a year at a million users, so a free tier is not a loss leader we tolerate — it is a permanent, sustainable position. That matters strategically because most of our users will never pay us anything by design: safety is never paywalled, and the people who need safety most are often the people least able to pay for it.

Put the other way: it is not that we refuse advertising and must therefore charge. It is that our cost structure is the only reason refusing advertising is survivable at all.

On advertising, and why refusing it costs us less than it looks

Advertising has the highest ceiling of any model in this industry and it is structurally unavailable to us. That is usually framed as a constraint. It is closer to the opposite.

Ad monetisation needs three things we cannot have by construction: identity resolution, so you know who this is across sessions; behavioural signal, so you know what they want; and attribution, so you know whether it worked. We hold none of the three. Untargeted inventory clears at a small fraction of targeted rates, so the honest statement is not that advertising is distasteful to us — it is that advertising would pay us almost nothing even if we were willing.

And there is a second-order benefit: our promise is credible because we cannot quietly renege on it. A company that could switch on ads tomorrow is a company that might. We would have to rebuild the product to do it.

What comes next, in order

The control that keeps this cheap at scale

In a product that cannot meter people, you must meter resources. The one place with a real marginal cost reachable without paying is relayed call bandwidth, because calls are free. We have modelled the abuse case properly — a rotating-identity user costs us tens of terabytes a day at scale, which would be genuinely existential — and the controls are enforced on our server rather than trusted to the app: each account gets a fixed number of relay credentials a day, each good for an hour, and every relay box refuses sessions beyond a hard bandwidth ceiling. What no control can stop is a user who makes a new identity for every call, which is why the ceiling is per box and not per person. We would rather you heard that from us.

Previous
The Team