Support

Where to write

Channel	What for
@zero_coo1	Support: API errors, payments, limits, integration questions
@MixenAI_bot	The main service bot
@MixenAPI_bot	The API bot: keys, balance, top-ups
@MixenAI	Channel: new models and changes

What to attach

The more of this your first message carries, the shorter the exchange:

• The response id — the id field in the response body, shaped chatcmpl-…; it identifies the request unambiguously.
• The request time with a timezone, and the model key (model).
• The error code and the full body — the error envelope carries type and code, which usually name the cause outright.
• The key prefix — the leading characters, shaped mxn-AbC3de, shown in your cabinet.

Never send the key itself. Not in a message, not in a screenshot: the prefix is enough to find it, whereas a leaked full key would have to be reissued.

What to check yourself first

• 401 — the key is wrong, revoked, or sent without Bearer.
• 402 — /v1 opens after the first top-up; see Errors, limits, billing.
• 403 — the key is valid but lacks the scope: an image key is not allowed into video.
• 404 on a model — the model was retired or renamed; the live list is always in GET /v1/models.
• 429 — the per-minute limit was exceeded; the response carries Retry-After.
• An empty answer with finish_reason: "length" — the whole max_tokens went into reasoning; see Reasoning effort.
• Balance and charges — GET /v1/balance and GET /v1/history.

Money

Refunds of an unused balance follow the refund terms. If a charge looks wrong, attach the response id and the time: the log keeps the actual spend of every request, and sorting it out takes minutes.