Questions and answers about ChatBoss

In ChatBoss a bot is assembled in a visual editor: you drag and connect ready-made blocks (message, buttons, AI answer, payment, appointment booking) without writing a single line of code. Each block is one step of the dialog, and the arrows between blocks define the order and the branching of the scenario. You can start from a blank canvas or from a ready-made template in the catalog. Once you press “Launch”, the bot is immediately available in Telegram.

A simple bot — a menu, lead collection, answers to frequent questions — takes 5–15 minutes if you start from a ready-made template. A bot with a complex AI dialog, appointment booking and payments usually takes from 30 minutes to a few hours, depending on the number of scenarios. There is no moderation stage: the bot goes live right after you build it.

The editor is built on a graph model: every block is a node and the links between nodes are transitions in the dialog, like in a flowchart. The “Condition” block branches the scenario by the user's answer (contains, equals, greater than and so on), so the bot can lead different clients down different paths. Changes are saved instantly and the scenario can be tested while you are still editing it.

Yes, both options live in the same editor. The template catalog contains ready-made scenarios for specific niches — salon appointments, order taking, customer support and others — which you can publish as they are or adapt. Alternatively you can start with an empty canvas and assemble the whole scenario from individual blocks yourself.

Everything collected through input blocks (name, phone, answers to questions, appointments) is stored in the “Tables” section of your account in a structured form — you can view and edit it right in the interface. The same data can be sent automatically to external systems through the HTTP block — to a CRM or Google Sheets, for example.

The number of bots is set by the account plan: the free plan runs one bot, Start allows up to 3, Pro up to 10. The plan is bought for the whole account rather than per bot: a single payment from $9 per month covers all of the owner's bots.

Publishing takes one click on “Launch” in the editor — the bot immediately becomes available at its t.me link and starts answering users. No extra verification or moderation from ChatBoss is required: you fully control the token and the settings of your own bot registered with @BotFather.

Yes, phones and tablets open the mobile editor: blocks are added from the bottom panel, links are made by tapping a block's output and then the target, settings open with a tap on the card; you can move and delete blocks, zoom with gestures and save. A bot can be built from scratch, but large schemes are easier on a computer, and creating or renaming block groups is only possible there for now.

The bot's system texts — payment and cart buttons, error messages — appear in the user's Telegram language: Russian, English, Portuguese, Indonesian or Spanish, and in Russian for other languages. Scenario texts are sent exactly as you wrote them, and the language of the AI block's replies is set in its settings. Template descriptions in the catalog are translated into the same five languages.

The AI block runs on OpenAI models: the current generation is GPT-5.6 (Luna, Terra, Sol), alongside GPT-5.2, GPT-5 Mini, GPT-5 Nano and the GPT-4.1 line. You pick the model for every AI block in its settings; the full list with prices is on the pricing page. The model affects cost and response speed: smaller models are cheaper and faster, larger ones are more accurate in complex scenarios.

The AI block is a graph node like any other: it receives the user's message and the conversation context, generates an answer through the selected model and can route the dialog into one of the branches depending on the user's intent. That lets you combine “hard” button-and-menu scenarios with a flexible AI dialog inside one bot.

Yes, there is the “Image Generation (GPT Image)” block on the gpt-image-2 model. It has two modes: “generate from text” creates a picture from a description, and “photo transformation” takes a photo sent by the user and draws a new image based on it. Formats are 1024×1024, landscape 1536×1024 and portrait 1024×1536, quality goes from Low to High. The block works from the Start plan and each image is paid from the balance; on the free plan you can place the block in the scheme, but that step is skipped.

The AI block can analyse images sent by the user — recognise a photo of a product or a document and answer based on its content, for example. Voice messages are transcribed into text, and the bot answers them like regular messages. Replies can be voiced too: the AI block settings have a “Respond with voice messages” option with a choice of voice.

Yes, the free plan ($0) includes the AI block on the gpt-4o-mini and gpt-5-mini models — enough to assemble a bot and put it to work. AI requests themselves are charged from the balance as they happen on every plan: 100 ₽ (≈ $1) is credited after you confirm your email. The free plan's limits are the “Created by ChatBoss.pro” signature in the reply to /start and unavailable image and video generation; larger models, media generation, removing the signature and higher limits require a paid plan from $9 per month per account.

Yes, the AI block keeps the dialog context for each user separately and takes previous messages into account when generating an answer, so the bot does not “forget” what you were talking about a minute ago. You can also store arbitrary data about the client through the “Session” block and use it elsewhere in the scenario.

Yes, the “Video Generation” block creates a clip from a text description through Leonardo.AI on Kling 2.5 Turbo or Veo 3.1 Fast. Length is 5 or 10 seconds on Kling and 4 to 8 seconds on Veo; the format is landscape, portrait or square (Veo has no square), and Veo adds sound. The block works from the Start plan, and the cost of the clip is charged to the balance before generation; model prices are on the pricing page.

Building a bot and signing up are free, and the first bot runs on the free plan for $0. There are two paid steps, both bought per account rather than per bot: Start — $9 per month, Pro — $19 per month. AI requests are paid separately and only as they happen: they are charged from the balance at the model prices, and no plan includes an AI quota.

Yes, the free plan costs $0 and is available to every account with no time limit — it is not a trial but a permanent option. It includes one bot, up to 100 active subscribers and 30 blocks in the scheme, the gpt-4o-mini and gpt-5-mini AI models, and in the reply to /start the bot adds a “Created by ChatBoss.pro” signature. The editor, templates and scenario blocks are fully available — except image and video generation: those blocks can be placed and saved, but they run from a paid plan.

The account balance is topped up in your dashboard — the current payment methods are shown in “Settings → Balance”. The minimum top-up is $1.18; exactly the amount you pay is credited to your balance.

Yes, that is what the “Payment” block is for. It accepts Telegram Stars right in the chat, TON or USDT on the TON network straight to your wallet, USDT on Tron (TRC-20), Alipay to your own merchant account, and Russian cards or SBP in rubles through Prodamus on your own payment page. The amount, the item name and the payment method are set in the block and can be filled in from scenario variables, and after payment the scenario continues by itself. How many payments a bot can accept per day depends on the plan.

No, the platform takes no commission on bot payments. Telegram Stars, TON and USDT, USDT on TRC-20, Alipay and Prodamus payments (card, SBP) go straight to the bot owner — to their bot, wallet, merchant account or payment page, not to a ChatBoss balance. Only the fee of the payment system or network itself applies.

Bots are not switched off: if there was not enough money to renew the plan, the account moves to the free plan ($0) and keeps working within its limits — basic AI models and the signature in the reply to /start. Whatever has already outgrown the free package (the number of subscribers, for example) keeps running in full for another 7 days, which is enough to top up without rushing. The bot's data — the scenario, the tables, the conversation history — is not deleted. AI requests are not executed at a zero balance: the AI block replies that the service is unavailable while the rest of the scenario keeps working.

Yes, 100 ₽ (≈ $1) is credited to the balance after you confirm your email — you can spend it on the bot's AI requests right away. Bonus funds are not counted when calculating referral commission, but they are fully spendable on the platform's services.

The AI cost is calculated by the tokens of the specific model: GPT-5 Nano is the cheapest one for simple answers, GPT-5.6 Sol is for complex consultations. You choose the model for every AI block in the scenario yourself, so you can balance answer quality against spending — fast models for simple branches and powerful ones only where it pays off.

The builder has booking built in: you set working slots by day of the week, the length of one appointment and the booking calendar — the client picks a free time right in the chat with the bot. The fastest way to launch such a bot is to start from a booking template in the catalog or from a ready-made appointment setup example.

Yes, the booking settings define working slots for each day of the week separately, and specific dates — holidays, a specialist's vacation — can be excluded from the calendar manually: on those days the bot will not offer an appointment.

That is configurable: you can allow a single appointment per slot — for one-on-one work with a specialist — or several parallel appointments at the same time, for a group class for example. The limit is set in the booking settings separately for each bot.

Yes, a reminder is built with the “Timer” block in the “At a date from a variable” mode. The timer takes the appointment date from the scenario, counts back the offset you need — a day or two hours before the visit, for example — and at that moment sends the client the next block with the reminder. A sending window keeps the message from arriving at night, and waits of up to 90 days are stored in a queue and survive a bot restart.

Yes, you add a choice of specialist or location to the scenario as a separate step before booking — through a menu block or an AI block — and then keep separate working slots for each option. The scenario is fully configurable.

The booking horizon is configurable — 90 days ahead by default, but the value can be changed to fit your business: limit booking to two weeks if the schedule changes often, or open several months for long-term planning.

Yes, that is the “Catalog” block: a storefront of products as cards or a list, with a photo, price and description. In cart mode the client picks several products and taps Checkout; in “buy now” mode they go to payment with a single product. The cart total goes to the “Payment” block, every checked-out order is saved in the Orders tab of the dashboard with the statuses Placed, Paid, Completed and Cancelled, and the “Notification” block sends your manager the order contents.

Yes. Recipients can be picked by segment: everyone, new in the last few days, active, long silent, or by language. A post goes out right away, at a set time or on a schedule — daily, weekly, monthly. AI can write the post from a short brief, you can send the post to yourself to check it before sending, and resend it to those who didn't receive it. The quota is delivered messages per month per bot: 1,000 on the free plan, 10,000 on Start and 50,000 on Pro.

There is the “Suggestion box” block for that: the user sends text, a photo, a video, a voice message or an album, and the bot copies it to the recipients without a “Forwarded from” label — the author stays anonymous. Each suggestion gets a number, and if you reply to it the bot passes your answer to the author, also anonymously. The block is not available in MAX: the author of a forward cannot be hidden there.

A giveaway is assembled from ready-made blocks. The “Random Number Generator” in list mode picks a prize with weights — that is how a roulette is made. “Calculate” awards points and can do it no more than once a day, “Leaderboard” shows the top participants and the user's own place, and “Channel subscription” only lets subscribers of your channels move on. The subscription check works in Telegram only.

The “Input”, “Button Menu” and “Payment” blocks have a Nudge setting: if the user hasn't answered within the time you set — from a minute to several days — the bot sends them your reminder message. There can be up to five such reminders per dialog, and a sending window keeps them from arriving at night. An answer after a reminder follows the normal branch of the scenario, and if the user moves to another step the reminder is cancelled.

Yes, but without a ready-made connector: CRM integrations are built with the “HTTP Request” block or the AI block in API mode. The block sends a request to any REST API with the headers, body and scenario variables you need, so the bot can create deals and leads in Bitrix24, AmoCRM or any other CRM with an open API — following that CRM's documentation.

There are no ready-made connectors to Google Sheets or 1C. The data — leads, contacts, bookings — is sent by the “HTTP Request” block or the AI block in API mode to a webhook or API endpoint that writes it to the spreadsheet or to 1C: for Google Sheets this is usually a small Google Apps Script, for 1C its HTTP service. To simply see the leads you don't need to connect anything: they are saved in the “Tables” section of the dashboard.

Yes, there is a dedicated SMS block supporting the SMSC.ru and SMS.ru providers and an arbitrary API — useful for duplicating an important notification to clients who have not opened the chat with the bot in Telegram yet.

That is what the “Operator” block is for — when it fires, the conversation with that particular user is switched to a live support agent, and further messages are handled manually instead of by the automatic scenario blocks.

When the scenario reaches the “Operator” block, the operator gets a Telegram notification — they can reply right there or in the Operator section of the dashboard, and the dialog is closed with a button or the /close command. The block has working hours and a break: outside the shift the client gets your message and the scenario moves on without waiting. The block works for clients in MAX too, and once the dialog is closed the user goes back to the scenario.

Yes, the dashboard has a broadcasts section for that: a post goes to all of the bot's subscribers or to a chosen segment — right away, at a set time or on a schedule. The “Notification” block is not meant for mass mailing: it sends a message to a fixed list of recipients, a manager for instance.

Every bot has an analytics page in the dashboard for 7, 30 or 90 days, with the change against the previous period: retention (how many people came back on another day), a funnel by blocks — at which step of the scenario users leave — and traffic sources by tagged links. There is also an AI dialog review: AI reads all of the bot's conversations and produces a review of them; the price is shown before you start and is charged to the balance.

Yes, the dashboard has a dialog history for every user. Each bot reply is labelled with the block that sent it, and a click opens that block in the editor. Problem spots — errors, a scenario that stopped, questions left unanswered — are highlighted, and one button takes you to them. The user list has search and filters, and any user can be banned: the bot will stop answering them, even on /start.

The bot sends the owner notifications: about new users (in a batch, at most once every two hours), about a new booking and an operator request — instantly, plus a weekly summary and a warning when activity drops. Each type is switched on separately, and the channel is Telegram, MAX or email; in automatic mode the first available one in that order is used. Important service notifications, such as an AI request blocked by the balance, always arrive.

Every bot has per-user limits on commands, button taps and messages within a short window; whoever keeps exceeding them is ignored for 5 minutes. AI blocks have a separate anti-fraud layer: a pause between requests and temporary bans of 5 minutes, an hour or a day for suspiciously frequent requests, doubled for repeat offenders. This protects the owner's balance from being drained.

No, bots are not switched off or archived. A bot on the free plan falls asleep after a day without messages but stays running: the very first message wakes it up and the reply arrives in 2–4 seconds, and in the dashboard the bot looks active the whole time. Bots on paid plans do not fall asleep.

By default messages are stored in the ChatBoss database, like on most SaaS platforms, and on top of that you can enable full end-to-end (E2E) encryption of the conversation for a particular bot — then even the ChatBoss team is technically unable to read the content of the dialogs.

It is AES-256-GCM encryption of the conversation: messages are encrypted before they are saved to the database, and decryption happens only in the memory of the bot's container on the fly so that the AI assistant can answer. It is switched on with the “Enable E2E encryption of conversations” option in the bot settings and is available to all users for free, at no extra charge.

Only the bot's owner — the encryption key is shown once when you enable the option and is also sent to Telegram and to your email, but it is not stored on ChatBoss servers. That means that even if the server is technically compromised, the content of the conversations cannot be recovered without that key.

E2E encryption in ChatBoss was built exactly for scenarios with sensitive data — medical consultations, financial and legal services, where requirements such as GDPR, HIPAA or local personal-data laws apply. That said, encryption is a technical safeguard and not a legal opinion about full compliance with a particular law: responsibility for processing personal data stays with the bot's owner as the data controller.

Without the key the already stored messages cannot be decrypted — ChatBoss does not keep the key on its side and cannot restore it. So it is worth saving the key in several places: it is sent both to Telegram and to your email at the moment encryption is enabled.

ChatBoss does not store payment details — card numbers or crypto wallet credentials: payments through Telegram Stars, TON, USDT, Alipay and Prodamus are processed by the payment systems and networks themselves and are not saved on the platform's servers. The password of a ChatBoss account is stored hashed, as industry security standards require.