💰Plans and limits

In Selarti you can choose two main plans — “Turnkey setup”, where our specialists will configure everything for you, and the “Self-setup” plan, where the cost depends on the number of AI managers (communication channels) and the volume of data in dialogues (depends on the chosen OpenAI language model). Understanding the limits is important so that your accounts are not blocked.

Cost of one AI manager — $20 per month

AI manager — a robot/bot based on artificial intelligence.

One manager can work on one task from one communication channel at a time. For example, if you need to work simultaneously with two Telegram accounts and one WhatsApp account, that is 3 managers.

Payment for the manager is charged automatically; if the balance is less than $20, the manager will be blocked until payment is made.

Volume of data in dialogues

When chatting with AI managers the cost or usage limit of the system is calculated based on the number of processed units of text or characters — tokens.

Token — the amount of text that can be processed. On average one word takes about 1-2 tokens for text in Russian, accordingly 1 token is roughly 3-4 characters.

Charges are applied only when the user responds.

Incoming and outgoing tokens

Incoming tokens — these are all the data sent to the model. They include the user's message, bot instructions, and chat history. The instructions account for the main prompt, references to the knowledge base and files, as well as requests from “Professional settings”.

Outgoing tokens — the text of the response that the model generates and sends to the user.

How a request is formed

One request consists of:

• the user's message — added to incoming tokens;

• bot instructions (prompt, knowledge base, “Professional settings”) — included in incoming tokens;

• chat history (context) — with each request the information is resent to the model and counted as incoming tokens;

• the model's response — consumes outgoing tokens.

The total cost of a request equals the sum of incoming and outgoing tokens.

Features of working with tokens

• Each new request is processed together with the chat history, so consumption increases in long dialogues.

• If the token limit is exceeded, the earliest part of the history is automatically deleted.

• Even in test mode responses are generated by the model, so charges from the balance also occur.

Example calculation

User message: “Yes, interested. Write what about the price” — 13 tokens. Bot instructions: “Answer politely, in simple language. Request contact” and functions — 101 tokens. Chat history: used so the bot understands the context — 395 tokens. Model response: information about prices and the catalog — 191 tokens.

Incoming = 13 + 101 + 395 = 509 tokens (at the rate of the chosen model). Outgoing = 191 tokens (at the rate of the chosen model).

Total cost 700 tokens (incoming tokens + outgoing tokens).

Incoming and outgoing tokens are billed separately. Usually incoming tokens are cheaper, and outgoing — more expensive.

Cost of operating different models

To choose between models you can use the article.

AI model	Incoming tokens (1M)	Outgoing tokens (1M)
gpt-5	$1.5	$12
gpt-5-mini	$0.3	$2.4
gpt-5-nano	$0.06	$0.48
gpt-4.1	$2.4	$9.6
gpt-4.1-mini	$0.48	$1.92
gpt-4.1-nano	$0.12	$0.48
gpt-4.5-preview	$90	$180
o1	$18	$72
o1-pro	$180	$720
o3	$2.4	$9.6
o3-deep-research	$12	$48
o3-pro	$24	$96
o3-mini	$1.32	$5.28
o4-mini	$1.32	$5.28
o4-mini-deep-research	$2.4	$9.6
gpt-4o	$3	$12
gpt-4o-mini	$0.18	$0.72
gpt-4-turbo	$12	$36
gpt-4-turbo-preview	$12	$36

Service “Turnkey setup”

Use the service for comprehensive configuration tailored to your business tasks.

To subscribe to the service — contact the sales department.

WhatsApp Business API (WABA) rates

To improve interaction between WhatsApp services and Selarti you can use the WhatsApp Business API integration.

WhatsApp Business API allows companies to integrate WhatsApp with Selarti for message automation, sending notifications, and customer support.

Rate $0.03 - $0.08 for opening a 24-hour conversation. The cost varies depending on the country. For Russia $0.085, rates for other countries you can see here.

Protection against bots.

If an AI manager encounters a bot in any of the channels, it will be blocked to avoid an endless dialogue. When blocked, the dialogue is marked as “bot” and “closed”. Incoming messages will arrive in the personal account, but the manager will not respond.

- maximum $1 per dialogue; - a maximum of 6 identical messages in a dialogue.

Message sending limits

To avoid blocks by services you must adhere to the limits.

There are no limits on processing incoming dialogues. The limit on outgoing communication is different for each channel. When creating a task with the work format “Incoming and outgoing communication” choose the limit of dialogues with users depending on the channel:

Channel type	Start	One week	Month	More than a month
Telegram Premium (Telegram cloud)	10	20	30	40
Vkontakte user	10	20	25	30
WhatsApp Web	5	5	7	10
WhatsApp Business API (WABA)	issued during registration by the WABA service: 200 and more	issued during registration by the WABA service: 200 and more	issued during registration by the WABA service: 200 and more	issued during registration by the WABA service: 200 and more
Vkontakte bot	issued during registration by the VK service: 200 and more	issued during registration by the VK service: 200 and more	issued during registration by the VK service: 200 and more	issued during registration by the VK service: 200 and more

For correct operation of the Telegram cloud channel it is recommended to enable Telegram Premium.

The “More than a month” column shows the maximum recommended daily limit of dialogues. Exceeding it risks blocks.

For new accounts (fresh registration or purchase on an exchange) these limits are the maximum. Established accounts that have been active for years can send more messages, but the risk of blocking still remains.

When configuring a task these limits should be specified in the “Outgoing limit per day per channel” field.