How it works Recraft V4?
Recraft V4 is an image generation model from the Recraft team. It accepts both text and images as input, allowing you to work from descriptions or references, with output at ~1K resolution in various aspect ratios. The frame composition is noticeably neater: objects and their placement more accurately match the intent. In BotHub, Recraft V4 is accessible without a VPN or foreign card — you pay with a Russian card in rubles only for what you use, and unused tokens do not expire. Alongside it in the same window are over 250 models, making it easy to switch and compare results. There is a unified OpenAI-compatible API for embedding generation into your services, AES-GCM encryption, and corporate access via contract with invoicing and an admin panel. Designers will find the model useful for concepts, posters, and covers; marketers for visuals for ads and social media in required proportions; product teams for illustrations for landing pages and presentations; and developers for generating images directly from code via API.Frequently asked questions about Recraft V4
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.
Recraft-v4 from Recraft generates images from text descriptions and references, and also works with text and documents. It is suitable for illustrations, banners, covers, social media graphics, variations of finished images, and describing uploaded images.
The model provides ready-made raster images that you can download directly from the BotHub interface. Adjust quality and detail with your prompt: specify style, composition, lighting, and materials. For a neat result, rely on a reference and refine the description in subsequent attempts.
A separate step-by-step reasoning mode is not noted in our data for recraft-v4, so you should not count on it. If you need a long logical chain before the result, switch to a model with confirmed reasoning capabilities in the same BotHub window.
Function calling and strict structured output are not noted in our data for recraft-v4. If this is critical for your scenario, use another model via the unified OpenAI-compatible BotHub API: you won't have to rewrite the integration, only the identifier changes.
Yes, images and documents are accepted as input: you can provide a reference to generate a new image or ask to describe an uploaded file. Audio and video are not in the list of supported input formats; for those, choose specialized models from the BotHub catalog.
There is no information about prompt languages in our data, so we cannot provide guarantees. In practice, a detailed description of the scene, style, and composition helps: if the result differs from your intent, rephrase the request or add a reference. Payment and access in BotHub from Russia without VPN.