← ClaudeAtlas

seedreamlisted

Generate, edit and split images into layers with ByteDance Seedream 5.0 (Pro, Flash, Lite) through the PopNow API: up to ten reference images, consistent characters and products, PSD-style layer decomposition. Exact USD quotes first. Use when the user asks for Seedream or multi-reference image editing.
popnow-ai/popnow-skills · ★ 0 · AI & Automation · score 75
Install: claude install-skill popnow-ai/popnow-skills
# Seedream 5.0 image generation, editing and layer splitting Seedream 5.0 is ByteDance's image family. Pro gives the most detail and accepts up to ten reference images, which keeps a character, product or outfit consistent across a set. Flash is the fast, cheaper tier, and Lite is the budget tier. Pro and Flash can also split a finished image into separate layers for editing. ## Models These are the PopNow model paths at the time of writing. Confirm them with `models` before use; newer versions appear there first. | Model path | Best for | |---|---| | `bytedance/seedream-5.0-pro/text-to-image` | Highest detail text-to-image | | `bytedance/seedream-5.0-pro/edit` | Edit with up to ten reference images | | `bytedance/seedream-5.0-pro/layer-decomposition` | Split an image into layers | | `bytedance/seedream-5.0-flash/text-to-image` | Fast, cheaper text-to-image | | `bytedance/seedream-5.0-lite/text-to-image` | Budget text-to-image | ## Prompt tips - Start drafts on Flash or Lite, then rerun the chosen prompt on Pro. - When using several references, name their roles in the prompt: "image 1 is the character, image 2 is the jacket, image 3 is the location". - For a series (storyboards, product angles, catalog shots), keep the same references and change only the scene line. - Layer decomposition returns separate layers; tell the user the outputs are links, not a single PSD file. ## Setup and cost Requires Python 3.8+ and a PopNow API key in `POPNOW_API_KEY`. Create a dedicate