The pricing we wished we had

Before Lana, we were customers. We ran a large agency and paid for chatting software like everyone else, and every invoice told the same story: a fee per seat for chatters who came and went, a fee per message that punished us for selling more, and a rate that never moved no matter how big we got.

When we built our own product, the pricing was not an afterthought. It was one of the first things we designed, with a simple brief: build the bill we always wanted to receive.

Two lines, nothing else

Lana's pricing fits on two lines. A licence per Telegram account, and a share of what Lana sells. That is the whole bill.

• The licence is $19 per Telegram account per month. Buy them one by one, or by pack for less: 10 for $170, 25 for $375, 50 for $650. Unused licences never expire, and moving one to another account keeps the days left.
• The share is 12% of the net sales Lana makes. It drops to 10% once your workspace passes $30,000 of net sales in a month, and to 8% above $100,000. Sales a person on your team sends by hand from Live Convs are at 4%, then 3%, then 2% on the same tiers, one counter for the whole workspace.

Net means net: what you actually receive after Telegram's share and refunds, never gross. One counter per workspace, all accounts together, reset on the 1st. Every sale in the console shows its rate and why.

There is no seat. Your whole team, from the owner to the night-shift chatter, works in the console at no extra cost. There is no message fee, no fee per creator, no fee for the API. You prepay a balance in USDT or USDC, licences and commission come out of it, and when it runs out Lana simply pauses until you top up. No contract.

Why the price can be this low

A share of sales only works if the AI behind it is cheap enough to run on every conversation, including the thousands that never lead to a sale. That is where most of our engineering time went.

Running a large language model on every fan message, with the creator's persona, Vault and scripts in context, is expensive if you do it the naive way. We did it that way first, at our own agency, then spent months bringing the cost down without touching the quality of the replies.

• Not every message needs the same model. A "hey" at 3 am and a fan negotiating a 1,500 Star video are not the same job. Lana routes each turn to the model that handles it best for the least, and reserves the heaviest reasoning for the moments that make or lose a sale.
• Context is built, not dumped. Instead of sending a creator's entire persona and history with every message, Lana sends what matters for this fan, this turn. Shorter prompts, faster replies, lower cost.
• Everything repeatable is cached. The parts of a conversation that do not change between turns are not paid for twice.
• We measure cost per sale, not cost per message. A cheaper reply that loses the fan is the most expensive reply there is, so every optimisation is checked against conversion before it ships.

The result is an AI that costs us a fraction of what a generic setup would, at the same or better quality in the chat. We pass that on. The $19 licence and the 12% share are not a promotional price: they are what the product costs to run well, plus a margin we are comfortable with.

Why the rate drops as you grow

Most tools charge big agencies more. More seats, more messages, enterprise tiers. We think that is backwards, for a reason that comes straight from running an agency.

The cost of serving an agency does not grow in line with its sales. Serving the first $30,000 of monthly sales means onboarding, setup, questions, tuning the personas, learning the market. Serving the next $70,000 on the same workspace costs us far less per dollar: the accounts are configured, the scripts are written, the AI Brain knows the creators. It would be dishonest to charge the same rate on both.

So the rate tiers down. 12% up to $30,000 of net sales a month, 10% from $30,000 to $100,000, 8% above. The same for manual sales: 4%, 3%, 2%. One counter, all accounts, so a multi-brand agency gets there faster than if each creator were counted alone.

There is a second reason, and it is the one we care about most: we want the agencies that grow on Lana to stay on Lana. A rate that drops as you grow means the product gets cheaper exactly when you have the most to lose by switching. That is the deal we wanted from our own tools and never got.

What this looks like in practice

Take an agency with 20 Telegram accounts doing $4,000 of net sales each per month, 80% of it sold by Lana. That is $80,000 of workspace net sales. Licences: two packs of 10, $340. Commission on Lana's sales: the first $30,000 of the workspace counter at 12%, the rest at 10%, blended on the AI share to about $6,880. Commission on the $16,000 sold by hand: about $380. Total around $7,600, or 9.5% of sales all-in, with no seat for the team.

The calculator on our pricing page does this maths live for your own numbers, and the AI Brain page shows what Lana actually does for that share.

We are in private beta and onboard agencies one by one. If you want the bill described above instead of the one you have, join our Telegram channel, where we announce each onboarding wave and how to apply.

FAQ

Is there a fee per chatter or per seat?

No. Your whole team works in the console at no extra cost. You pay a licence per Telegram account and a share of what Lana sells, nothing per person.

Why is the commission on manual sales so much lower?

Because Lana did not do the work. When a person on your team sends the paid offer from Live Convs, the sale is billed at 4%, 3% or 2% depending on your monthly tier, instead of 12%, 10% or 8%.

Does the rate reset every month?

Yes. The counter of workspace net sales resets on the 1st. Within the month, every sale is charged at the rate of the tier the counter is in at that moment, and the console shows the rate on each sale.