搜索你想找的开源项目

输入名称、用途或关键词,也可粘贴 GitHub 仓库链接
全部技能

skilltoAI/skillto-ai-studio

Skillto.ai|个人工作台

  • 19 个技能
  • 139 Star

装得最多的 5 个

  • Route Skillto-AI-Studio requests spanning content collection, media analysis, visual workflows, visual-style selection, talking-head editing, covers, tables, reports, and manuscript rewriting.

  • Write image-node prompts for an unpredictable new subject's styled look portrait or multi-view character card, with optional costume and prop transfer from a source-frame reference.

  • Acquire a source short video, produce frame-evidenced scene and per-second character breakdowns, then write a concise new-character replacement prompt grounded in the original performance and camera timeline.

  • Turn an authorized short-video URL into a local depth-map video, extracted audio, frame-evidenced shot document, and a SkillTo.ai depth-gray replacement workflow through Creator MCP.

  • Build or repair local Python crawlers for Douyin and Xiaohongshu account/video data when the user provides authorized cookies, logged-in sessions, account URLs, or asks for SQLite/web analysis of post metrics.

其余 14 个技能
  • 为AI知识、科普讲解、新闻与行业解读类短视频规划以完播和互动为目标的画面:对标视频拆解、证据清单、节拍图、视觉系统、分镜表、关键帧与发布前审片。口播剪辑和纯图文生成都适用,在 talking-head-edit 或其他生成流程之前使用。

  • Turn an authorized TikTok or Douyin short-video URL into an editable SkillTo.ai H3 character-transfer Workflow that prepares unpredictable user photos as character references before video generation.

  • Collect short-video media for the Skillto-AI-Studio character and scene store, preserve each source character's visual style and recognizable prototype in source character cards on exact

  • Ingest Motion template catalogs, links and .motion packages into the visible Skillto-AI-Studio library using a validated metadata, facet, vector and media contract; retrieve templates and produce videos through supported adapters.

  • Use the LAN AMD-PAD Qwen3.8 service for multimodal image or video understanding, content classification, frame extraction, shot-by-shot breakdowns, OCR, and structured media analysis. Use when local media must be inspected or classified; do not use for downloading protected media or ordinary text-only questions.

  • Create, inspect, update, archive, and restore local SkillTo Metabase reports and their charts through the 127.0.0.1:8880 report service.

  • Research source posts or videos, analyze manuscript rhetoric and video keyframes, and maintain source-backed second-creation drafts and versioned storyboards in the local 洗稿 workspace with human five-heart review.

  • 检索、比较和按规范录入可拆解美术风格,提供元素分层、独立入场/idle/出场与 Remotion/Motion 实现选型。用于风格选型、参考图归类、风格库维护和设计交接;色系与内容保持为项目变量,保留白模等形式的固有特征。

  • Build and maintain local or AMD-PAD remote multimedia tables organized by knowledge base and project, with shared report storage, human review pages and authenticated semantic MCP operations for AI.

  • Turn rough-cut 口播 A-roll, timecoded SRT, and optional B-roll into a publish-ready, editable video with HyperFrames planning, Remotion editing/rendering, and video-talkcraft motion recipes.

  • Collect authorized TikTok profile and keyword-search video data, expand multilingual pet-content keywords, and analyze recent popularity or low-follower viral content with SQLite and report-table exports.

  • Install and use the Skillto-AI-Studio Chrome popup to download authorized video pages or MP4/HLS/DASH resources into a persistent local video breakdown workspace.

  • Create, inspect, edit, validate, archive, and run SkillTo.ai visual workflows through Creator MCP, including node prompts, model options, media inputs, and result retrieval.

  • 为“小与封面图”生成统一品牌的视频、课程和AI教程封面。支持调用skillto-ai-studio-douyin-xhs-crawler拉取指定视频,分析场景关键帧与ASR,制作XR/AR沉浸式或多案例对比式封面。固定IP垫图,IP必须手持或操作主题相关专业道具,绿白粗笔刷字、黑金电影感,imagegen默认生成3:4与4:3两版。