Skip to main content
Use Seedream 5.5 Pro through AnyFast for text-to-image generation, reference-image editing, and transparent image output. The request model ID is seedream-5-5-pro. ByteDance’s Seedream 5.5 announcement describes the series’ multi-turn editing, native RGBA, realism, and text-rendering capabilities.
To request transparent output, set background to transparent and output_format to png. A request without background: "transparent" returns an opaque image.

Key capabilities

  • Text to image — Create images from natural-language prompts with realistic textures and clear rendered text
  • Image editing — Refine a reference image over multiple turns while preserving regions outside the requested edit
  • Transparent background — Generate images with full or partial transparency in native RGBA output

Generate from text

The Seedream 5.5 series is designed to render lighting, materials, and skin textures naturally and to improve text rendering. Describe the intended composition, materials, lighting, and any exact text in the prompt.

Output controls

Size and input limits

The size options 1K, 1.5K, and 2K were accepted. Their returned dimensions were 1248x832, 1888x1248, and 2512x1680, respectively. The response includes the actual dimensions in data[].size. For image editing, pass a source image URL in image. See the official Seedream 5.5 overview and Seedream 5.5 announcement for model capabilities.

API Reference

View all Seedream 5.5 Pro workflows.