[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"pack-detail-content-creator-ai-studio-zh":3,"seo:pack:content-creator-ai-studio:zh":97},{"code":4,"message":5,"data":6},200,"操作成功",{"pack":7},{"slug":8,"icon":9,"tone":10,"status":11,"status_label":12,"title":13,"description":14,"items":15,"install_cmd":96},"content-creator-ai-studio","🎬","#F43F5E","new","本周新建","内容创作者的 AI 工作室","给 YouTuber、播客主、Newsletter 作者、TikTok 创作者的十件 AI 资产：选题、剧本、配音（云端 + 开源）、字幕、缩略图、B-roll 生成、加上一个能把这些串起来的发布平台。",[16,28,36,44,51,58,66,73,81,89],{"id":17,"uuid":18,"slug":19,"title":20,"description":21,"author_name":22,"view_count":23,"vote_count":24,"lang_type":25,"type":26,"type_label":27},4297,"721d23c5-ffea-448c-b2a6-67c905855aad","claude-code-agent-content-marketer-721d23c5","Claude Code Agent: Content Marketer","Use this agent when you need to develop comprehensive content strategies, create SEO-optimized marketing content, or execute multi-channel content campaigns to drive engagement...","TokRepo精选",46,0,"en","skill","Skill",{"id":29,"uuid":30,"slug":31,"title":32,"description":33,"author_name":34,"view_count":35,"vote_count":24,"lang_type":25,"type":26,"type_label":27},2978,"ad11ab44-64e2-4ece-9d8a-2a072fda98e3","elevenlabs-voice-design-generate-voices-from-prompts","ElevenLabs Voice Design — Generate Voices from Prompts","ElevenLabs Voice Design generates new voices from text prompts. Describe age, accent, tone — get a voice you own and reuse via TTS API.","ElevenLabs",102,{"id":37,"uuid":38,"slug":39,"title":40,"description":41,"author_name":42,"view_count":43,"vote_count":24,"lang_type":25,"type":26,"type_label":27},2462,"e7a8aaaf-453a-11f1-9bc6-00163e2b0d79","styletts-2-human-level-text-speech-via-style-diffusion-e7a8aaaf","StyleTTS 2 — Human-Level Text-to-Speech via Style Diffusion","A TTS system that achieves human-level speech synthesis through style diffusion and adversarial training with large speech language models. Fast inference with natural prosody.","Script Depot",108,{"id":45,"uuid":46,"slug":47,"title":48,"description":49,"author_name":42,"view_count":50,"vote_count":24,"lang_type":25,"type":26,"type_label":27},390,"e1fd7c46-bbda-4956-8649-9c3ed579ff25","whisper-cpp-local-speech-text-pure-c-c-e1fd7c46","whisper.cpp — Local Speech-to-Text in Pure C\u002FC++","High-performance port of OpenAI Whisper in C\u002FC++. No Python, no GPU required. Runs on CPU, Apple Silicon, CUDA, and even Raspberry Pi. Real-time transcription.",1602,{"id":52,"uuid":53,"slug":54,"title":55,"description":56,"author_name":42,"view_count":57,"vote_count":24,"lang_type":25,"type":26,"type_label":27},2458,"7e2317bb-453a-11f1-9bc6-00163e2b0d79","cogvideo-text-image-video-generation-7e2317bb","CogVideo — Text and Image to Video Generation","An open-source video generation framework from Zhipu AI supporting text-to-video and image-to-video with CogVideoX models. Generates high-quality clips up to 6 seconds.",155,{"id":59,"uuid":60,"slug":61,"title":62,"description":63,"author_name":64,"view_count":65,"vote_count":24,"lang_type":25,"type":26,"type_label":27},776,"84500559-5ce6-41c7-ba22-9712153bb821","together-ai-image-generation-skill-claude-code-84500559","Together AI Image Generation Skill for Claude Code","Skill that teaches Claude Code Together AI's image generation API. Covers FLUX and Kontext models for text-to-image, image editing, and style transfer with correct parameters.","Together AI",125,{"id":67,"uuid":68,"slug":69,"title":70,"description":71,"author_name":42,"view_count":72,"vote_count":24,"lang_type":25,"type":26,"type_label":27},2348,"044138c3-43a5-11f1-9bc6-00163e2b0d79","imagemagick-command-line-image-processing-200-formats-044138c3","ImageMagick — Command-Line Image Processing for 200+ Formats","ImageMagick is a free, open-source software suite for creating, editing, compositing, and converting images. It supports over 200 image formats including PNG, JPEG, TIFF, WebP, SVG, and PDF.",162,{"id":74,"uuid":75,"slug":76,"title":77,"description":78,"author_name":79,"view_count":80,"vote_count":24,"lang_type":25,"type":26,"type_label":27},101,"7775f06a-8adf-477a-91e9-85f51682cd10","remotion-captions-subtitles-ai-powered-video-subtitles-7775f06a","Remotion Captions & Subtitles — AI-Powered Video Subtitles","AI skill for generating and rendering captions in Remotion videos. Supports transcription, word-level timing, and styled subtitle export.","Skill Factory",197,{"id":82,"uuid":83,"slug":84,"title":85,"description":86,"author_name":87,"view_count":88,"vote_count":24,"lang_type":25,"type":26,"type_label":27},1328,"300e919c-381e-11f1-9bc6-00163e2b0d79","ghost-professional-publishing-platform-modern-journalism-300e919c","Ghost — Professional Publishing Platform for Modern Journalism","Ghost is an open-source publishing platform built for professional publishers. It bundles a blazing-fast Node.js CMS, Substack-style paid memberships, email newsletters, and SEO — everything a modern publication needs, self-hosted.","AI Open Source",194,{"id":90,"uuid":91,"slug":92,"title":93,"description":94,"author_name":42,"view_count":95,"vote_count":24,"lang_type":25,"type":26,"type_label":27},2434,"05ad6f38-44f8-11f1-9bc6-00163e2b0d79","yt-dlp-feature-rich-audio-video-downloader-05ad6f38","yt-dlp — Feature-Rich Audio & Video Downloader","yt-dlp is a feature-rich command-line tool for downloading audio and video from thousands of websites. A community-maintained fork of youtube-dl with active development, format selection, post-processing, and SponsorBlock integration.",106,"tokrepo install pack\u002Fcontent-creator-ai-studio",{"pageType":98,"pageKey":8,"locale":99,"title":100,"metaDescription":101,"h1":102,"tldr":103,"bodyMarkdown":104,"faq":105,"schema":121,"internalLinks":127,"citations":140,"wordCount":153,"generatedAt":154},"pack","zh","内容创作者的 AI 工作室 —— YouTube\u002F播客\u002FNewsletter 10 件套","Content Marketer agent、ElevenLabs Voice Design、StyleTTS 2、whisper.cpp、CogVideo、Together AI 出图、ImageMagick、Remotion 字幕、Ghost、yt-dlp —— 一个独立创作者用来跑「选题→剧本→配音→视频→缩略图→分发」全流程的 10 件资产。从 TokRepo 安装。","内容创作者的 AI 工作室 —— 独立创作者的制作团队","十件资产，按生产流水线排好：先选题与剧本，再做配音（云端+开源兜底），然后字幕、B-roll 生成、缩略图，最后用一个发布平台同时打通博客与 Newsletter。专为请不起剪辑的 YouTuber\u002F播客主\u002FNewsletter 作者准备。","## 这个 pack 里装了什么\n\n这是一个独立创作者意识到「兼职剪辑回不来了」之后会自己搭起来的工具栈。十件资产覆盖**内容流水线的每一个环节** —— 从「这周该做什么选题」到「邮件发出去了、视频上传了、缩略图渲染好了」。每一件都在真实生产流程里有明确分工。\n\npack 刻意在两个最容易让 API 账单失控的环节 —— **配音和图像生成** —— 同时给了云端方案和开源兜底方案。摸索内容形态时先用云端版本，等每周产量把账单上的那一行刷得让你不舒服时，再切到自托管版本。\n\n它默认你是个**多平台创作者**：同一份剧本会变成一支视频、一集播客、一条 Newsletter 短文，再加三条 TikTok 切片。yt-dlp + whisper.cpp + Ghost 这三件二次创作工具的意义就是：让周二的一份剧本产出周二的五条内容。\n\n## 安装顺序（选题 → 剧本 → 制作 → 剪辑 → 分发）\n\n1. **Claude Code Agent: Content Marketer** —— 从这里开始，因为选题不对其他都白搭。一个 Claude Code 子 agent，把一句话点子展开成 outline、hook、节奏点，并生成不同平台版本（长视频脚本 + 60 秒短切 + 推文 thread）。\n2. **ElevenLabs Voice Design** —— 目前最不像「机器人念稿」的云端 TTS。用 Voice Design 通过 prompt 铸造一个属于你自己的稳定旁白，把那个 voice ID 长期复用，让你的频道有个能被认出来的「声音标识」。\n3. **StyleTTS 2** —— 当 ElevenLabs 账单超过你心理价位时切到的开源 TTS。基于 style-diffusion，质量接近人声，单张消费级 GPU 就能跑。作为兜底，或者拿来配「不需要是『那个声音』」的 B-track 旁白。\n4. **whisper.cpp** —— 本地 STT。这是默默承担最多脏活的工具：把原始录音转成文字方便剪辑、生成字幕、喂二次创作流水线（文字稿 → Newsletter 短文 → 推文 thread）。完全离线跑，未剪辑 B-roll 不出本机。\n5. **CogVideo** —— 文+图生短视频，专门给你没实拍素材的 B-roll 兜底。6 秒一段，够覆盖「办公桌前的人」「城市街景」这类填充镜头。它不替代真实素材，它替代的是 stock 库订阅。\n6. **Together AI Image Generation** —— 缩略图与频道艺术图引擎。托管的 Flux\u002FSD 模型 + 干净 API + 独立开发者负担得起的定价。一分钟出 8 张缩略图候选，挑最好的发出去。\n7. **ImageMagick** —— 命令行图像处理。「把这张缩略图分别 resize 到 1280×720、1080×1080、1920×1080」这类活，一个平台一行 bash 搞定。是这个 pack 里你会用得最多但最不会去吹的工具。\n8. **Remotion Captions & Subtitles** —— 给短视频烧字幕。TikTok \u002F Shorts \u002F Reels 用户默认静音看，字幕不是「加分项」是「必备项」。Remotion 字幕渲染把样式写成代码，让每个视频排版一致。\n9. **Ghost** —— 既是发布平台也是 Newsletter 引擎。一个真相源：长文 + 发给订阅者的邮件 + SEO 友好的公开页面，三件事同一份内容。一个可自托管的 Node 应用替代「WordPress + Mailchimp + Buffer」。\n10. **yt-dlp** —— 二次创作流水线的基石。下载自己过去的节目、嘉宾的旧演讲、想剪片的会议录像 —— 同一条命令搞定所有平台。下游接 whisper.cpp 转写、接 Remotion 剪片。\n\n## 它们怎么串起来（ASCII 内容流水线）\n\n```\n           ┌── Content Marketer Agent ──┐\n           │ （点子 → outline → 剧本）   │\n           └──────────────┬──────────────┘\n                          ▼\n          ┌── ElevenLabs Voice Design ──┐\n          │   或 StyleTTS 2（开源）      │\n          │   （剧本 → 旁白 WAV）        │\n          └──────────────┬───────────────┘\n                          ▼\n     ┌──── whisper.cpp（转写旁白） ────┐\n     │            ▼                     │\n     │      SRT + 纯文本                │\n     │       │            │              │\n     │       ▼            ▼              │\n     │   字幕         Newsletter 草稿   │\n     │  （Remotion）  （Ghost）         │\n     │       │            │              │\n     │       ▼            ▼              │\n     │   B-roll       订阅邮件          │\n     │  （CogVideo）   + 公开文章 URL    │\n     └────────┬─────────────────────────┘\n              ▼\n  ┌── Together AI 图像生成 ──┐\n  │   （缩略图候选）          │\n  │            │              │\n  │            ▼              │\n  │     ImageMagick           │\n  │  （resize 1280×720 \u002F      │\n  │   1080×1080 \u002F 1920×1080） │\n  └───────────────────────────┘\n              │\n              ▼\n        yt-dlp（之后）\n        把发出去的视频拉回来 →\n        切片 → 二次创作\n```\n\n两个关键拼接点：**whisper.cpp → 字幕 + Newsletter**（同一份文字稿喂两个出口）、**Together AI → ImageMagick**（一张生成的缩略图变成三个平台尺寸）。把这两个点拼对，单条内容的人力消耗直接减半。\n\n## 你会遇到的取舍（DIY AI 声音 vs 真人，AI 缩略图 vs 设计师）\n\n- **DIY AI 声音 vs 真人配音** —— 2026 年 AI 声音越过了「不让人尴尬」这条线，但还没越过「听起来像一个真的在乎这个话题的人」那条线。频道门面的旁白人格请自己录。播客里的广告口播、不能露脸的平台、B-roll 旁白用 AI 配，10 倍速出片。先用 ElevenLabs Voice Design 摸清哪部分该真人、哪部分该 AI。\n- **AI 缩略图 vs 设计师** —— Together AI 60 秒给你「能用」的缩略图。真人设计师 4-8 小时给你「点得动」的缩略图。在内容增长期保持周更，AI 缩略图加 5 分钟人工调整（文字叠加、裁剪、对比度）在速度上完胜外包。粉丝过 10 万、每提升 1 个 CTR 点都值真金白银时，再请设计师。\n- **ElevenLabs 费用 vs StyleTTS 2 自托管** —— ElevenLabs 独立开发者档 $22-99\u002F月，音质确实更好。StyleTTS 2 在你已有的 GPU 上免费跑，音质*差不多*够。经验法则：每周生成音频 30 分钟以内留在 ElevenLabs；超过这个量自托管的账自然平。\n- **Ghost vs Substack\u002FBeehiiv** —— Ghost 是「自托管、自己拿订阅者名单」的开源方案：你拥有名单，平台不会某周二改规则。代价：你得维护一台服务器。Substack 是「租平台流量」的方案：零运维，但他们想改条款随时改。如果你已经在管自己的网站，选 Ghost。如果你下周二就要发刊，选 Substack。\n- **CogVideo vs 付费 stock 素材** —— CogVideo 适合做填充式的场景空镜。在观众**盯着看**那段镜头时（特写、人脸、特定动作）就不行了。给真正承担信息的镜头预算 $20-40\u002F月买 Pexels Pro \u002F Artgrid 这类正经 stock 库，CogVideo 用来做切镜过渡。\n\n## 常见踩坑（听起来像 AI、版权陷阱）\n\n- **视频开始有「AI 味」**。症状：每篇剧本都用「让我们深入了解」「在这期视频里我们将探索」「XX 的世界」。原因：模型默认表达渗进了你的口吻。解法：维护一份 `style.md` 让 Content Marketer agent 每次先读 —— 明确列出「禁用短语」「句子节奏示例」「你的犀利观点是 X、Y、Z」。每月更新一次。\n- **生成声音\u002F图像的版权陷阱**。ElevenLabs Voice Design 生成的声音是你的、可商用，但**未经授权**克隆他人的声音是禁区、会被平台封号。Together AI 出图同理：多数模型允许商用，但训练数据来源不够干净，别生成「模仿某在世艺术家风格」。\n- **字幕烧错宽高比**。16:9 渲染字幕再裁切成 9:16 给 Shorts，右半边字直接没了。一定要按最终平台宽高比直接渲染；Remotion Captions skill 有显式的 `width`\u002F`height` 参数 —— 用上。\n- **忘了按平台重编码**。YouTube 要 H.264 + AAC 高码率；TikTok 偏好稍低码率 + 激进 web 优化。同一份源文件，不同导出。给每个平台加一步 ffmpeg 处理；别上传一份 master MP4 然后听天由命。\n- **Newsletter 和视频对不齐**。周二发视频，周五 Newsletter 引用它，结果你忘了把 Ghost 里缩略图换掉。解法：一篇内容对应一篇 Ghost 文章，邮件从那篇文章**生成**。别并行起两份草稿。\n- **二次创作只复制不改写**。把视频文字稿原封不动贴进 Newsletter 是最懒的操作，读起来也确实是。让 Content Marketer agent 把文字稿**改写**成 Newsletter 语气（短句、不要「就像我视频里说的」、换个新 hook）。复用观点不复用句子。",[106,109,112,115,118],{"q":107,"a":108},"10 件都要装吗？还是可以先上小一点的子集？","先上四件：Content Marketer agent 出剧本、ElevenLabs Voice Design 出旁白、whisper.cpp 出文字稿+字幕、Ghost 出博客+Newsletter。这四件就能跑起一个完整的 YouTube + Newsletter 流水线。手做缩略图烦了再加 Together AI + ImageMagick；开始发 Shorts\u002FReels 再加 Remotion Captions；需要 B-roll 再加 CogVideo；ElevenLabs 账单开始扎眼再加 StyleTTS 2；决定要把老节目剪二创时再加 yt-dlp。10 件全装只有在每周出多条内容时才有意义。",{"q":110,"a":111},"对一个独立创作者来说，这套每月实际花多少？","现实基线（每周 1 视频 + 1 Newsletter）：Hetzner $5\u002F月 跑 Ghost，whisper.cpp \u002F StyleTTS 2 \u002F ImageMagick \u002F yt-dlp 全开源 $0，ElevenLabs starter $22\u002F月，Together AI 出图独立开发者用量 ~$5\u002F月，CogVideo API 按需 $0-10\u002F月。合计大约 $40\u002F月，加你的 Claude 或 GPT 订阅给 Content Marketer agent 用。涨得最快的成本线是 ElevenLabs；StyleTTS 2 存在的意义就是替换它。",{"q":113,"a":114},"AI 生成的旁白会让我 YouTube 频道被取消盈利吗？","单独用不会。YouTube 在 2026 年的立场：AI 内容只要有清晰的创意输入、不是「批量、重复、低质」就可以盈利。你写的剧本、你剪的视频、你定的编辑观点，加一个 AI 旁白，不会触线。会被砍盈利的是：20 个频道上传同一份 AI 生成剧本、同一个 AI 声音、同一段 AI B-roll。你的品味才是护城河。",{"q":116,"a":117},"为什么是 Ghost 不是直接 Substack？","Substack 上手更快 —— 注册、写、发。Ghost 需要你跑一台服务器（或者付 Ghost Pro $9-25\u002F月）。那为什么还选 Ghost：（1）订阅者名单完全归你，平台不在中间，（2）它是真正的 CMS，同一篇内容直接成为 SEO 友好的公开页面，不只是个邮件存档，（3）平台不能某天改分成或改内容政策来卡你。如果你打算做 5 年以上选 Ghost，如果你下周二就要发刊选 Substack。",{"q":119,"a":120},"字幕、配音、B-roll 真的能从一份剧本同时生成吗？","能 —— 整条流水线就是为这个设计的。Content Marketer agent 出剧本。剧本送进 ElevenLabs（或 StyleTTS 2）生成旁白 WAV。WAV 送进 whisper.cpp 生成带时间戳的 SRT（通过 Remotion 变成字幕）。同一份剧本还会作为分镜 prompt 送进 CogVideo 生成 B-roll 片段。周二一份剧本，自动旁白、自动字幕、自动 B-roll 建议。你仍然在 loop 里把关品味 —— 选最好的 take、修不对劲的 B-roll —— 但人工转写、对时、找素材这些活全没了。",{"@context":122,"@type":123,"name":13,"description":124,"numberOfItems":125,"inLanguage":126},"https:\u002F\u002Fschema.org","ItemList","十件 AI 资产，专为独立创作者（YouTube \u002F 播客 \u002F Newsletter \u002F TikTok）准备：选题、剧本、配音、字幕、缩略图、B-roll、发布。",10,"zh-CN",[128,132,136],{"url":129,"anchor":130,"reason":131},"\u002Fzh\u002Fpacks\u002Fvideo-production-ai","AI 视频生产线 pack","流水线扩展到要批量出片时，Remotion\u002FMoviePy\u002FOpenMontage 覆盖装配线那一侧",{"url":133,"anchor":134,"reason":135},"\u002Fzh\u002Fpacks\u002Fvoice-ai-stack","语音 AI 栈 pack","实时语音 agent（LiveKit \u002F Cartesia \u002F Moshi）把这里的 TTS 选型延伸到现场播客与语音对话",{"url":137,"anchor":138,"reason":139},"\u002Fzh\u002Ffeatured","TokRepo 精选资产","这 10 件资产生活在更大的 agent-ready 创作者工具目录里",[141,145,149],{"claim":142,"source_name":143,"source_url":144},"ElevenLabs Voice Design 通过文字 prompt 生成自定义声音并保留可复用的 voice ID","ElevenLabs Voice Design 官方文档","https:\u002F\u002Felevenlabs.io\u002Fdocs\u002Fproduct-guides\u002Fvoices\u002Fvoice-design",{"claim":146,"source_name":147,"source_url":148},"whisper.cpp 是 OpenAI Whisper 的高性能 C\u002FC++ 移植版本，可完全本地运行且无依赖","ggerganov\u002Fwhisper.cpp on GitHub","https:\u002F\u002Fgithub.com\u002Fggerganov\u002Fwhisper.cpp",{"claim":150,"source_name":151,"source_url":152},"Ghost 是专为专业出版者打造的开源发布平台，内置 Newsletter 引擎","Ghost 官方网站","https:\u002F\u002Fghost.org\u002F",1720,"2026-05-22T07:30:00Z"]