PRODUCT INTRO VIDEO SCRIPT · v2
Vita — Product Intro Video Script
v2 改动说明(相对 v1)
-
开场合并为一个 Capture 场景(原 v1 的 Hook + Capture 两段合并成一段):
- 拿掉 "three months later" 的时间设定——三个月听起来太久,容易让人觉得"这东西没那么重要"。
- 不再强调"记得它存在过、但找不到",改为强调准确保存 + 吃透理解——海量内容看过就能精准存下来,每个细节都记得。
- 补齐多模态输入的具体例子:播客、YouTube(重点)、Slack 消息、Twitter/X。
- 新增"口喷"场景:想法通过口头说出来(不是打字/不是正式记录)也能被捕获下来。
- "Ancher eats the real world" 这句英文口播换掉——直译"吃下世界"在英文口播里不够准确/不够地道,v2 改成更具体的表达(见下方逐字文案)。
- Recall 场景改写为"精准召回一句话":不再是"整篇文章/整段播客被召回",改成一个具体痛点场景——一个 45 分钟的 YouTube 视频,真正有用的只有其中 10 秒的一句话,Ancher 能精准把那 10 秒连同出处一起找回来,而不是把整段视频扔给你。
- Living Context / Identity-Grounded Output / Ship-Ready Multi-Format Creation / Flywheel / 收尾五段文案未改动,只是时间轴因为前两段变长而顺延。
使用说明(给拍摄/剪辑同事)
- 左列时间轴仅供参考,按实际配音和剪辑节奏微调。
- 右列"画面提示"是方向性建议,不是唯一答案——镜头语言可以按你的判断调整,但不要新增或替换未在本文档出现的功能描述。
- 全片没有出现任何具体数字(如接口数、用户数)——这些暂未确认,剪辑时不要自行加字幕补数字。
- CTA 里的链接/按钮文案是占位符,正式版发布前需替换成当前官网真实链接。
脚本
| 时间 | 口播(VO) | 画面提示 |
|---|---|---|
| 0:00–0:22 | A podcast episode. A YouTube video. A Slack thread. A tweet. Even the idea you just spoke out loud on your way to a meeting. Ancher captures all of it — not just the file, the substance. Every detail, saved with precision, actually understood. | 多模态输入蒙太奇:播客波形卡片、YouTube 视频卡片(画面占比要大,作为重点素材)、Slack 消息气泡、X/Twitter 卡片,依次汇入统一的知识结构;新增一个"口喷"镜头——人物在走路/通勤途中对着手机说话,一段语音波形从嘴部/麦克风图标涌出,同样被吸入同一个结构 |
| 0:22–0:45 | Notes wait for you to file them. Ancher's context doesn't. It organizes and connects itself around the goals you're actually chasing — growing every month, like a compounding asset, not another pile to clean up later. | 界面动效:零散卡片自动围绕一个"目标"节点聚拢、连线成图;对比一个杂乱堆叠的传统笔记文件夹画面 |
| 0:45–1:12 | That YouTube video ran forty-five minutes. What actually mattered was ten seconds — one line, from one creator, buried two-thirds of the way through. Ancher doesn't just recall the video. It recalls that exact line — the moment you actually need it, with the source attached. | 一条很长的视频进度条(大部分为中性灰色,代表 45 分钟),其中一小段(约 10 秒)高亮为金色并轻微脉动;该片段从进度条中"抽出",落入用户当前正在写的文档/工作台,附带创作者名字和时间戳的引用标注 |
| 1:12–1:37 | Every model can write well. Only Ancher writes like you — your voice, your taste, your standards, your role. It learns from what you save, what you linger on, what you rewrite, what you approve. Not a generic answer. Yours. | 对比画面:左边是"通用 AI 输出"(灰色、平淡的文本),右边是"Ancher 输出"(带有用户风格标记、被采纳打勾);可用真实产品截图 |
| 1:37–2:00 | And the output isn't a wall of text in a chat window. It's ship-ready — HTML pages, decks, videos, podcasts, reports, complete with citations, ready to send. One body of context. Every format you need. | 画面:一份 context 素材,"分裂"成多个成品格式的动效(deck、HTML 页、报告并排出现),最终是一个可以直接发送/发布的画面 |
| 2:00–2:17 | Capture. Context. Recall. Identity. Creation. Every output flows back in — so the longer you use it, the more it becomes unmistakably yours. | 画面:五个能力节点首尾相连成一个循环动画(flywheel),中心文字浮现:"Judgment compounds." |
| 2:17–2:37 | Ancher — the AI workspace where judgment compounds. [CTA 占位:Try it at ancher.ai / 官网真实链接] | Logo 定版,CTA 按钮/链接画面,联系方式或试用入口 |
与基准的偏离说明(校对用)
功能基准 landing page 的原始一句话仍然是权威定义,以下两条只是这条视频的口播措辞做了改写,能力定义本身没有变:
- Capture 原句:"Others need clean files. Ancher eats the real world — one memory that travels across every platform and every model." → v2 口播不再直译 "eats the real world",改成 0:00–0:22 那段更具体的描述(播客/YouTube/Slack/Twitter/口喷 + "actually understood")。如果以后要做纯 tagline 类的短版素材(比如 15 秒版本),"eats the real world" 这句原句仍然可以用,只是这条 2:35 长版口播不用它。
- Task-Time Recall 原句:"Search finds documents. Recall brings back what you already knew." → v2 口播换成 0:45–1:12 那段具体的 YouTube 45 分钟/10 秒场景,是同一个能力的举例化表达,不是新能力。
未改动的三条(原句照旧)
- Living Context:"Notes wait for you to file them. Context organizes itself around your goals."
- Identity-Grounded Output:"Every model can write well. Only Ancher writes as you — your voice, your taste, your standards, your role."
- Ship-Ready Multi-Format Creation:"Tools make one format from scratch. Ancher makes every format from everything you know."
来源:Ancher_5_core_capabilities_v1_zh.md + 已确认的 landing page。
拍摄前必须补的信息(本稿留白)
- CTA 最终链接/按钮文案
- 是否需要真人出镜(创始人/团队)开场或收尾,还是全程屏幕录屏 + 口播
- 配乐风格(无提供方向,建议下属先出一版试配乐再定)
- "口喷"镜头具体怎么拍:真人实拍 or 图形化表现(麦克风+波形图标),待定