360 Secure SkillHub
360 安全技能中心

Poyo Seedream 4 5

该工具用于PoYo平台的Seedream4.5图像生成与编辑,核心支持标准生成、图像编辑,可输出2K/4K高分辨率图,支持最多10张参考图及1-15张图输出。典型场景为用户明确指定Seedream4.5,或有高分辨率生成、图生图、多参考编辑需求。输入需prompt(编辑需额外提供image_urls),可设置尺寸、输出数量;输出含模型ID、参数摘要、任务ID等,指引后续状态查询。

coolhackboy
coolhackboy
热度 60.27
最近更新 2026-07-09 10:54:41
分类:创作依赖:第三方 API / 代码执行 / 读写文件应用场景:图像生成

安装技能

复制下面的话术发给你的 Agent,它会通过 secure-skills 检索并安装该技能。

执行命令安装此 skill: npx secure-skills-cli add poyo-seedream-4-5
Metadata
name
poyo-seedream-4-5
description
Seedream 4.5 image generation and editing on PoYo / poyo.ai via `https://api.poyo.ai/api/generate/submit`; use for `seedream-4.5`, `seedream-4.5-edit`, 2K/4K output, multi-reference editing, and higher image counts.
metadata
{"openclaw":{"homepage":"https://poyo.ai/models/seedream-4-5-api","requires":{"bins":["curl"],"env":["POYO_API_KEY"]},"primaryEnv":"POYO_API_KEY"}}

PoYo Seedream 4.5 Image Generation and Editing

PoYo Links

Use this skill for Seedream 4.5 jobs on PoYo. It covers high-resolution generation, multi-reference image workflows, and edit requests.

Use When

  • The user explicitly asks for Seedream 4.5, seedream-4.5, or seedream-4.5-edit.
  • The task is high-resolution generation, image-to-image, or edit.
  • The workflow needs multiple reference images or higher output counts.

Model Selection

  • seedream-4.5: standard generation entry point.
  • seedream-4.5-edit: use when the request explicitly edits supplied images.

Key Inputs

  • prompt is required.
  • image_urls supports up to 10 images and is required for seedream-4.5-edit.
  • size can be a resolution token like 2K/4K or an aspect ratio like 1:1, 16:9, 21:9.
  • n controls output count from 1 to 15.

Execution

  • Read references/api.md for endpoint details, model ids, key fields, example payloads, and polling notes.
  • Use scripts/submit_seedream_4_5.sh to submit a raw JSON payload from the shell.
  • If the user only needs a curl example, adapt the example from references/api.md instead of rewriting from scratch.
  • After submission, report the task_id clearly so follow-up polling is easy.

Output expectations

When helping with this model family, include:

  • chosen model id
  • final payload or a concise parameter summary
  • whether reference images are involved
  • returned task_id if a request was actually submitted
  • next step: poll status or wait for webhook