← Back to home
Skill

pavo-drama — AI 短剧生产管线

tencent_skillhub · 0.9.8 · Downloads 519 · Heat 46.9

Introduction

Generate short-drama videos end-to-end via the Agnes AI API (a.k.a. PavoAPI, served by Sapiens AI). Two-stage pipeline: Stage 1 — Asset Pipeline: generate character / scene / prop key visuals via agnes-image-2.1-flash, audit them with a TWO-GATE system — G1 file-level hard inspection (format / dimensions / square / face-ratio, see scripts/asset_inspector.py) then G2 multimodal scoring (resolution, defects, style consistency, card compliance) using agnes-2.5-flash multimodal understanding, retry failures, catalog approved assets to assets/index.json with public URLs. Characters are generated as a 6-pose asset pack (front_anchor / side_45 / full_body_riding / 3 expressions) at 1024x1024 square with bilingual bust-composition prompts. Then render an ASSET-REVIEW HTML gallery (assets/review.html) for human sign-off BEFORE any video budget is spent. Stage 2 — Episode Production (v0.9.0: FOUR STRUCTURAL GATES + CROSS-SHOT FACE LOCK): writer agent produces a shot table referencing asset IDs (and, when given source_doc/source_facts/beats, writes FROM the source novel); consistency-checker gates era/forbidden terms; the STORYBOARD gate compiles + validates the ToonFlow 六铁律 and aborts the whole episode on any error; a continuity BASELINE locks cross-shot staging + a story state machine; shot-director binds assets as i2i image conditions, INJECTS locked CharacterDNA reference images into every frame render, runs a FACE FIDELITY GATE (multimodal comparison) immediately before submitting, and COLOR GRADES all clips for lighting/tone consistency — so a drifted/hallucinated shot is never sent to the video API. Passing shots render as FIRST+LAST FRAME (mode=keyframes) videos via agnes-video-v2.0 with async polling; the final short-drama ASSEMBLY is deferred (off by default) and done separately via `assemble-episode`. Shares the SAME Agnes backend that ToonFlow (AI 短剧工厂) wraps via its local HTTP API — see "ToonFlow Integration" below for driving short-drama production through ToonFlow's orchestrator/DB instead of (or alongside) calling Agnes directly. Cross-shot / cross-episode consistency is maintained by character cards, scene cards, prop cards, fixed seeds per asset, and asset reuse via the catalog. Load this skill whenever the user asks to produce short dramas, 短剧, 微短剧, 竖屏短剧, or vertical short-form video content via PavoAPI / Agnes AI / ToonFlow.

Resource Information

Author:user_cc65e9bc
Category:Multimedia
Version:0.9.8
Auth: Required
Pricing: Not specifiedAPI key required

License information is collected from official sources and may be outdated or missing. Refer to the LICENSE file in the original repository before commercial use.

Tags
图片生成视频处理

Information boundary

Information on this page is collected from official documents for technical presentation only and does not imply our endorsement. Pricing, quotas and availability are subject to the latest official statements. We make no commitment regarding any action you take after leaving this site.

Similar skills

视频短视频图片音乐音频配音口播数字人全能创作

Marked paid

通过同一个AI创作工具完成AI图片、视频、音乐与语音创作,查找公开社交数据,编辑视觉成果,并集中查看和管理生成资产。Beatra 全能创作支持AI图片生成(文生图)、AI视频生成(图生视频)和AI视频编辑,通过AI音乐生成把想法做成歌曲或纯音乐,也可用AI语音生成(文字转语音)、多语言旁白和AI语音克隆完成跨媒介内容;还能查找抖音数据、小红书数据,以及 YouTube、TikTok、Instagram 与 X 上的公开帖子、笔记、评论、账号和趋势,把社交媒体数据用于竞品调研和创作。它是一款面向AI内容创作与多模态内容创作的AI创作套件,用起来就是一个把图片、视频、音乐与语音收在同一处的AI创作助手,适用于社交内容、产品展示、营销素材、课程、播客、短片和跨媒介项目;还能在不同媒介间复用原始素材、已有成果和查到的公开资料,查看制作进度、返回文件、实际用量和积分记录。

AI数字人口播与虚拟主播视频

Marked paid

用一张人像和一段短脚本或语音,制作有明确讲述内容的数字人口播视频。可使用现成录音,也可以先选择音色,把脚本生成旁白,让画面中的人物随声音完成稳定、自然的口播。适合 AI 主播、虚拟主播视频、数字人讲解、让照片说话、照片说话视频、产品讲解视频、数字人培训视频、公告和社交口播,并在交付后检查人物一致性、声音清晰度、口型同步和动作稳定性。

罗斯蒙德可视化图形生成

Not specified

【罗斯蒙德可视化图形生成】南京罗斯蒙德基于已交付项目沉淀的可视化图形生成场景:根据业务描述或系统需求,自动生成可编辑的架构图、流程图、ER 图、时序图、状态图与 C4 图。采用「统一图模型」一次建模、多格式输出:结构化文本(Mermaid / PlantUML,可嵌入 Markdown、进 Git 版本管理)、可编辑图形文件(drawio / Excalidraw,可拖拽修改节点与样式)、静态图(SVG 自渲染,PNG 给转换命令),并输出图形说明(图的含义 + 关键节点解释)。内置样式系统统一配色与命名,内置布局引擎自动排布,校验脚本拦截孤立节点、悬挂连线与语义缺漏。触发词:画架构图、画流程图、画 ER 图、画时序图、画状态图、C4 图、系统架构图、部署架构图、数据流程图、实体关系图、生成 drawio、生成 Excalidraw、导出 mermaid、导出 PlantUML、导出 SVG、导出 PNG、可编辑的图、图转可编辑、架构图可编辑、流程图二次修改、画图、出图、画个图。

架构图一键生成

Marked paidAPI key required

使用 ContextWeave 生成或修改架构图、流程图、思维导图和复杂信息图。适用于需要从代码、文件或自然语言中提取结构与关系并实际产出 CW、SVG、HTML、可编辑 PPTX 或原生 VSDX 的请求;未特别说明时,PPTX 与 Visio 使用原生形状和连接器导出。不适用于统计图表、手绘插画或像素级排版。

AI Image Prompts · 绘画提示词库

Marked free

Recommend curated prompts from a 10,000+ real-world image generation prompt library. Works with ANY AI image model — Nano Banana Pro, Nano Banana 2, Seedream 5.0, GPT Image 1.5, Midjourney, DALL-E 3, Flux, Stable Diffusion, and more. Use this skill when users want to: - Find proven image generation prompts (any model) - Get prompt inspiration for portraits, products, social media, posters, etc. - Create illustrations for articles, videos, podcasts, or marketing content - Browse categorized prompt templates with sample images - Translate and understand prompt techniques

架构图一键生成

Marked paidAPI key required

使用 ContextWeave 生成或修改架构图、流程图、思维导图和复杂信息图。适用于需要从代码、文件或自然语言中提取结构与关系并实际产出 CW、SVG、HTML、可编辑 PPTX 或原生 VSDX 的请求;未特别说明时,PPTX 与 Visio 使用原生形状和连接器导出。不适用于统计图表、手绘插画或像素级排版。

What is a Skill

A Skill is a packaged single capability, such as text recognition, document summarization, or keyword extraction. Developers wrap the calling logic and prompts into a plugin-like unit, so users can load it without understanding the underlying model. StackPilot indexes a large number of skills that can be tried instantly or assembled into a plan.