Simplicity
A three-step flow: photo, look or description, result. Complex models hide under clear buttons.
We make AI photo and video generation simple: ready-made looks, your face, results in Telegram — no editor and no long settings.
MagicBox is a platform for generating images and short videos with neural nets. The service combines several models (photo, video, faceswap) in one bot and Mini App: you pick a look or describe a scene, upload a photo — the model does the rest.
We start from the idea that people shouldn’t need prompt syntax to get a frame “with themselves”. A catalog of looks and clear scenarios matter more than a panel of 40 sliders.
Unlike “pure” generators where you write a request from scratch every time, MagicBox offers ready scenes and a “upload photo → get result” flow. Prompts remain for those who need them.
Models and pipelines change: we connect what gives the best balance of face quality, speed, and credit cost.
Give users clear access to current AI photo and video models in one place — with a transparent credit price before generation starts.
Принципы, на которые опираемся в продукте
A three-step flow: photo, look or description, result. Complex models hide under clear buttons.
Priority — recognizing yourself in the frame. Separate pipelines for that: faceswap, identity, Pro models.
Inside the service everything is counted in credits. The cost of a specific generation is visible before confirmation.
Photos are needed to build the frame. We don’t use them longer than required to deliver the result and run the service.
Ключевые этапы — заготовка, даты уточните перед продом
First photo and video generation scenarios in Telegram, provider tuning and face quality.
Looks catalog, credit licensing model, public site pages, and legal framework.
Вопросы по сервису и сотрудничеству
Новости, поддержка и сам продукт — в мессенджере.
Для юридических и деловых обращений.
Первые генерации можно сделать в боте. Карта для теста не обязательна, если включён бесплатный лимит.