Token导航 LogoToken导航TokenDH.com
开发执行命令github未标认证来源可访问许可证需确认审计提醒

ui2v-apiui2v API 文档

Agent Skill

用于辅助 API 设计、接口文档、请求响应结构和服务集成说明。它适合让 Agent 梳理 endpoint、生成 OpenAPI 草稿、检查字段命名、整理错误码或辅助前后端联调。使用时需要确认真实业务语义、鉴权方式、分页和错误处理规则;涉及生成接口文档时,应避免凭空补字段,最好从现有代码、schema 或接口样例中提取事实。

总安装

569

周安装

23

GitHub Stars

公开资料未说明

下载量

178
CodexClaudeCursorGemini CLI

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

2

许可证

unknown

最后核验

2026-05-01

来源状态

来源可访问

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:ui2v-api(ui2v API 文档)
来源仓库:https://github.com/illli-studio/skills
仓库路径:skills/ui2v-api
安装命令:
npx skills add https://github.com/illli-studio/skills --skill ui2v-api
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/illli-studio/skills --skill ui2v-api

简介

用于辅助 API 设计、接口文档和请求响应结构梳理。

  • 适合生成 OpenAPI 草稿、检查字段命名或整理错误码。
  • 使用时需确认业务语义、鉴权方式和分页规则。
  • 避免凭空补字段,应从现有代码或接口样例中提取事实。
  • 建议结合前后端联调需求提供服务集成说明。ui2v-api 属于开发类 Skill,可作为该场景下的辅助能力补充。

SKILL.md

UI2V HTTP API Skill

Use the local UI2V server at http://127.0.0.1:5125.

Use this skill for

  • Text-to-video generation
  • Text-to-poster generation
  • Custom dimensions
  • Optional styles
  • Async job submission, polling, and result download on this workstation
  • Reusable PowerShell scripts instead of ad hoc one-off terminal code

Output types

UI2V supports two primary generation modes:

  • Video generation: animated output such as mp4 or webm
  • Poster generation: static output such as png or jpg

Both flows can behave asynchronously on this workstation. Do not assume posters are always synchronous just because they are images.

Use this workflow

  1. Ensure UI2V is reachable at http://127.0.0.1:5125.
  2. Treat GET / returning 404 as acceptable. On this machine that still means the service is up.
  3. Submit either POST /video or POST /poster with the correct payload.
  4. Inspect the submit response:

- If the content type is video/* or image/*, save the stream directly. - If the response body is JSON and contains requestId, switch to async polling.

  1. Poll GET /status/:requestId until the job reaches completed or failed.
  2. When status becomes completed, download GET /result/:requestId to the final output path with the expected extension.
  3. If status becomes failed, report the full status payload and retry only the failed job.

Use the bundled resources

  • scripts/ui2v-common.ps1: Shared submit, poll, download, and liveness helpers used by both entry scripts.
  • scripts/invoke-ui2v-video.ps1: Thin entry script for one video request.
  • scripts/invoke-ui2v-poster.ps1: Thin entry script for one poster request.
  • references/async-media-flow.md: Detailed async response shapes, endpoints, and operational notes.

Prefer the scripts over ad hoc terminal one-liners whenever the task involves actual delivery work.

Request shapes

Video request body:

{
  "prompt": "Educational motion graphic with floating equations",
  "format": "mp4",
  "width": 1080,
  "height": 1080,
  "quality": "medium",
  "style": "optional-style-name"
}

Poster request body:

{
  "prompt": "Educational poster with books, globe, and bold title area",
  "format": "png",
  "width": 1080,
  "height": 1080,
  "style": "optional-style-name"
}

Quality tiers

  • low: fastest preview quality
  • medium: default balance for day-to-day generation
  • high: stronger final quality
  • ultra: slower, higher fidelity
  • cinema: slowest, premium video output when available

Only video requests use quality. Poster requests do not require it.

Handle statuses deliberately

  • queued: The task is accepted and waiting. Keep polling.
  • generating: Rendering or export is still in progress. Keep polling.
  • completed: Download from /result/:requestId.
  • failed: Stop, capture the error payload, and decide whether to resubmit.

Work with both media types safely

  • If an .mp4 file is only a few hundred bytes, it is probably JSON job metadata that was mistakenly written straight to disk. Open the file, parse the requestId, then poll and download the real result.
  • If a .png or .jpg file is only a few hundred bytes, handle it the same way. It may contain queued-job JSON rather than image bytes.
  • If the status payload contains Export already in progress for this project, too many jobs reached export at the same time. Re-run the failed request after another export finishes, or serialize the batch.
  • If the API returns 429 Busy, wait and retry rather than changing prompts or dimensions.
  • If polling times out, show the last status JSON in the final response so the user can see what stalled.

Batch safely

For multiple videos or posters, prefer sequential execution. Submitting several jobs at once can work, but downloading and exporting them in parallel is less reliable on this machine.

Video batch example:

$jobs = @(
  @{ Prompt = "Math lesson motion graphics"; Output = ".\\math.mp4" },
  @{ Prompt = "Science lab motion graphics"; Output = ".\\science.mp4" }
)

foreach ($job in $jobs) {
  powershell -ExecutionPolicy Bypass -File .\skills\ui2v-api\scripts\invoke-ui2v-video.ps1 `
    -Prompt $job.Prompt `
    -OutputFile $job.Output `
    -Width 1080 `
    -Height 1080 `
    -Quality medium `
    -Overwrite
}

Poster batch example:

$jobs = @(
  @{ Prompt = "Back to school poster"; Output = ".\\poster-1.png" },
  @{ Prompt = "STEM workshop poster"; Output = ".\\poster-2.png" }
)

foreach ($job in $jobs) {
  powershell -ExecutionPolicy Bypass -File .\skills\ui2v-api\scripts\invoke-ui2v-poster.ps1 `
    -Prompt $job.Prompt `
    -OutputFile $job.Output `
    -Width 1080 `
    -Height 1080 `
    -Format png `
    -Overwrite
}

Start UI2V if needed

If the API is not reachable, start the desktop app first.

Start-Process "C:\Program Files\UI2V\UI2V.exe"

Poster note

Poster generation is a first-class capability of this skill, not a footnote. Keep its request format, output extension, and downstream usage distinct from video generation, but handle its async job lifecycle with the same care.

适合场景

01

用户想查找某类 Agent Skill 时

02

需要根据任务场景推荐可安装能力包时

03

需要对比不同来源的安装命令和来源信息时

能力概览

能力 1

按任务关键词查找相关 Skills

能力 2

展示可复制的安装命令

能力 3

保留来源站点、仓库和原始说明,方便继续核验

能力 4

展示第三方安全扫描或审计结果

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

Codex

35.64%
按下载量换算63

Claude

30.73%
按下载量换算55

Cursor

17.04%
按下载量换算30

Gemini CLI

9.53%
按下载量换算17

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

可疑

权限和风险

执行命令

安装流程涉及命令执行,可能通过 npx skills add https://github.com/illli-studio/skills --skill ui2v-api 联网下载 Skill 或依赖。用户安装前应确认命令来源、仓库内容和执行环境。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。来源安全扫描存在 warning/failed 结果,不能写成本站确认安全。当前只有一个来源,正式发布前建议补源仓库或其他目录站核验。

来源信息

继续浏览同类 Skills