Pixyn FAQ: tokens, payment, generation

By card. Sole proprietors and companies can pay by invoice from a bank account. Plans and packs are on the subscriptions page.

Yes. A fiscal receipt is generated automatically after a successful payment and sent to the email you entered at checkout. This is required by Federal Law 54-FZ: a receipt is issued for every payment. If there is no receipt, check your spam folder and that the address is correct.

Tokens are usually credited within one or two minutes. Refresh the page or check your balance in the Telegram bot; make sure the payment is completed at the bank and not still "processing"; check that the payment and the balance belong to the same account. If more than 10 minutes have passed, write to an operator with the payment time and the last 4 digits of the card. You do not need to send the full card number, support never asks for it.

Yes. On the payment page choose the business option, fill in your company details, and we will issue an invoice; closing documents (an act or a UPD) are sent to the email you provide. Tokens are credited after the money reaches our bank account, so not instantly: if you need access right away, paying by card is faster. More on the business page.

Tokens are the internal currency of Pixyn; every generation is paid with them: an image, a video, music, text. The cost depends on the neural network and the settings: the heavier the task, the more tokens it takes. Tokens are charged when a generation starts. If a generation fails because of the system, the tokens are returned to your balance automatically.

Tokens from the monthly subscription pack do not carry over to the next period, keep this in mind when choosing a plan. Tokens from purchased packs never expire and stay after the subscription ends.

No, it does not renew on its own: we do not store your card and never charge anything without you. When the paid period ends, the account moves to the free plan, there is nothing to cancel or turn off. To continue or change your plan, pay for the period you need on the subscriptions page.

Refund terms are described in section 7 of the public offer. Refunds are handled by a support operator; there is no automatic refund button, so write to the chat or to contact@pixyn.ru.

Common reasons: the request was rejected by the content policy of the neural network; the input file is larger than the model limit; the neural network is overloaded or timed out; the format and size settings are incompatible. What to do: rephrase the request, make the file smaller, wait a couple of minutes, or reset the settings to default. Tokens for failures caused by the system are returned automatically. If the same error keeps repeating, send its text to an operator.

Neural networks reject requests with famous people, explicit content, violent scenes and protected characters. This is a restriction of the models themselves, it cannot be bypassed. It helps to replace the name of a real person with a description of the role ("a famous actor", "a businessman in a suit") and to soften the description of the scene. You can upload photos of real people for editing if the content itself is allowed.

It depends on the neural network you choose: limits in the catalog range from 14 to 48 MB. If the file is larger, the generation will not start, and the system will show the exact maximum for that model. It helps to compress the image (JPG instead of PNG) and reduce the resolution to 2K or 4K; this barely affects the result, because neural networks bring the input to their working resolution anyway.

Images take from a few seconds to a couple of minutes, videos from one to ten minutes depending on the model, duration and load. At peak hours the queue is longer, that is normal. You can close the page: the generation finishes on the server, the result appears in your history, and the Telegram bot sends a notification. If a task hangs for more than 30 minutes without a result or an error, write to an operator.

Both. On the website the main mode is the studio: choose a category and a neural network, write a request, attach references. In the Telegram bot the same generations are available from the chat, together with your balance and ready notifications. The account and the balance are shared.

There is no universal "best" neural network: some follow the request more precisely, others give cinematic video, others handle text in images or faces better. Where to look: the model catalog with descriptions and examples, and the knowledge base with ready prompts. A practical tip: on a new task, make one test generation in two or three models before running a series.

Yes. Issue a key in your account: "Profile → API access". The key gives your code the same neural networks and settings as the studio and charges tokens to the same balance. There is no separate plan for access. The API documentation describes how to connect and how to run generations from Claude and ChatGPT via MCP. Series of similar generations are automated inside the platform with the visual WorkFlow: the pipeline is set up once and runs in one click.

The customer. Under clause 9.1 of the public offer, all exclusive rights to the results you create while using the platform belong to you; Pixyn may not use them without your written consent, except in cases directly provided for by the contract or the law.

From 16, this is clause 13.2 of the public offer. By accepting the offer, you confirm that you have reached this age.

Write a request to an operator in the support chat or to contact@pixyn.ru from the address your account is registered to. The account, data and created materials are deleted within the period set by law. What we store and why is described in the privacy policy.