recraft-ai / recraft-v3

Recraft V3 (code-named red_panda) is a text-to-image model with the ability to generate long texts, and images in a wide list of styles. As of today, it is SOTA in image generation, proven by the Text-to-Image Benchmark by Artificial Analysis

  • Public
  • 352K runs
  • Paper
  • License

Pricing

This model is priced by how many images are generated.

Check out our docs for more information about how per-image pricing works on Replicate.

Readme

Overview

Recraft introduces a revolutionary AI model that thinks in design language. This new model has set a new standard for excellence in image generation and follows a period of rapid growth for the graphic design platform. Since its launch in 2023, Recraft has been used by over 1.5 million users in more than 200 countries.

Model Features

  • Improved image generation quality through superior prompt understanding, leading to more accurate visual representations of the designer’s intent. As a result, designers gain enhanced visual aesthetics, improving anatomy and overall image quality. Image generation quality of the latest model is better than all existing image generation models proven by public benchmarks
  • Accurate text generation in images. Recraft’s new model is the first to offer image generation with text of any size and length. This feature sets it apart from others — while very few models can offer similar capabilities, they still can’t match Recraft’s ability to generate long texts within an image.
  • A unique graphic design generator that enables seamless and precise integration of text and image elements into AI-generated designs. It is now possible to set the exact positions of text and different elements in the image instead of interpreting the explanation in the prompt to get the needed result - just drag and drop images into frames and use the model’s controlled AI generation to ensure their vision comes to life strictly as intended.
  • Brand style customization feature allows users to upload reference images for the AI to generate branded content. The new style creation flow allows for fine-grained experimentation with style details to get exactly the style required for the brand’s unique look and feel.
  • Recraft’s model supports both raster and vector image generation. It stands out from AI generators like MidJourney and OpenAI because their models don’t support the ability to generate vector art. Apart from that, Recraft V3 supports an extensive list of AI image editing tools.

Team Expertise

The team behind Recraft brings extensive machine-learning expertise and experience to the graphic design industry. Founder Anna Veronika Dorogush is the creator of CatBoost, one of the most popular Machine Learning frameworks in the world, with hundreds of thousands of weekly downloads.

Licensing and commercial use

If you generate images on Replicate with Recraft-v3 models, then you can use the images commercially.