How it works Recraft V4 Styles?
Recraft V4 Styles is an image generation model by Recraft, built around working with references: each request includes at least one style sample image, and the model creates a new picture, replicating its drawing style. This approach keeps the visual language of a series consistent when many images are needed that must look like a single collection. In BotHub, the model is accessible from Russia without a VPN or foreign card, payment is in rubles based on actual usage, and you can switch from Recraft to any of over 250 other models in the same window. A unified API allows connecting generation to your services, and for companies, there is a contract, invoice, and admin panel with limits. It is useful for designers needing a series of illustrations in one style, marketers for banners and creatives under a single brand guide, authors of presentations and articles, product teams for icons and interface design, and studios assembling sets of visuals based on an approved sample.Frequently asked questions about Recraft V4 Styles
You can use generated results for commercial purposes. You own all rights to the content you create. The only restriction: make sure your prompt does not include copyrighted third-party material. You are responsible for respecting the rights to any input data.
This is a generative Recraft model for images: it creates images from text descriptions and processes uploaded ones based on a reference. Suitable for illustrations, banners, icons, covers, product cards, and maintaining a consistent visual style for a brand. Works in one BotHub window without a VPN.
Exact resolutions and file formats are not specified in our data — refer to the technical metrics at the top of the page. The model's focus is in its name: neat stylization, clean composition, readable graphics. A context window of up to 65,536 tokens allows for detailed scene and style descriptions.
A separate step-by-step reasoning mode is not noted in our data, and it is usually not required for an image generator: the model processes the prompt in one pass. If you need to think through an idea first, do it in a BotHub text model, then pass the finished description to recraft-v4-styles.
Function calling and strict structured output are not noted for the model itself in our data, so you shouldn't rely on them in your scenario. However, it is accessed via the unified OpenAI-compatible BotHub API: the call is easy to integrate into code, and you can replace the model if needed without rewriting the integration.
Images — yes, this is image-to-image mode: you provide a reference image and a text edit, and the model redraws it in the desired style while maintaining composition recognizability. Audio and video inputs are not noted in our data; for such files, choose a specialized model in BotHub.
There is no information about supported prompt languages in our data, so we cannot provide guarantees. The easiest way to check is through practice: the BotHub interface is in Russian, payment is via Russian cards and based on actual token usage without a subscription, so a couple of test generations will show the result immediately.