Skip to content

armgit / hypit

Public Imported from GitHub
0

Clone any viral video with AI agents. Not just a script, the whole workflow: swap the face, the words, the B-roll, ship 100 variants in one command, and get your 100M views.

docs: refine agent quickstart flow

rponeee98aaaf316e
diff --git a/docs/quickstart.md b/docs/quickstart.md
index 862941fb..5d16f084 100644
--- a/docs/quickstart.md
+++ b/docs/quickstart.md
@@ -20,36 +20,24 @@ after you understand the Agent workflow below.
- A Coding Agent that can use skills, such as Claude Code or Codex.
- A folder where the Agent can create your video project.
- A reference video or your creative brief, if you are making an original video.
-- Credentials for the models and services you choose. The Agent asks for these only when they are
- required by the selected Runtime Profile.
-You do not need to clone the Hypit repository. The Agent can install the reusable `/hypit` skill and
-the Hypit Distribution for you. When a setup command is necessary, the Agent explains what it does
-and runs it in the project directory.
+You do not need to clone the Hypit repository. Install the reusable `/hypit` skill with the command
+in Step 1; the Agent can then prepare the Hypit Distribution for you. When a setup command is
+necessary, the Agent explains what it does and runs it in the project directory.
-## 1. Open the Hypit skill and provide credentials
+## 1. Install the Hypit skill
-Start a conversation with your Coding Agent in the folder where you want the project to live. Ask it
-to use Hypit and sign in:
+Start a conversation with your Coding Agent in the folder where you want the project to live. Install
+the Hypit skill once with:
-```text
-/hypit login
-```
-
-The skill guides you through the login flow and stores credentials using the configured credential
-store. You may instead tell the Agent to use your own provider keys. For example, you can say:
-“Use my KIE and Gemini keys; do not create another account.” The Agent will identify which keys the
-requested video actually needs, explain why each one is needed, and avoid asking for unrelated
-credentials. Never paste a secret into a public document or commit it to Git.
-
-If the skill is not installed yet, ask the Agent:
-
-```text
-Install the Hypit skill, then log me in and check that my environment is ready.
+```bash
+npx skills add hypit-ai/hypit -g
```
-The Agent performs the installation and the local environment checks. You do not have to memorize
-package managers, Runtime Profiles or service-specific setup instructions.
+The `-g` option makes the skill available to new Agent sessions in any project folder. You do not
+need to clone the Hypit repository or understand the files this command installs. Once the skill is
+available, start a new Agent conversation in your video project folder and continue with the steps
+below.
## 2. Describe the video you want
@@ -61,15 +49,14 @@ Ask the Agent to clone a video and provide the file path. You can drag the video
Coding Agent window; most agents insert its path for you. You can also write the path explicitly:
```text
-/hypit Clone this video and preserve its pacing, shot structure, captions, B-roll rhythm and sound design:
-/Users/me/Desktop/reference.mp4
+/hypit clone this video: /path/to/video.mp4
```
Tell the Agent what should change, if anything: the host, language, product, aspect ratio, visual
style or call to action. The reference is treated as evidence about the editorial structure. The
Agent does not merely summarize it; it creates an editable program that can be reviewed and built.
-### Create an original video from a brief
+### Create an original video
You can describe an idea just as you would brief a human producer. For example:
@@ -84,7 +71,23 @@ colors, presenter, platform or aspect ratio. You do not need to decide which pac
caption system or animation primitive implements those choices. The Agent turns the brief into a
complete plan and asks follow-up questions only when a creative decision cannot be inferred safely.
-## 3. Let the Agent do the production work
+## 3. Provide credentials when the Agent asks
+
+After you describe the video, the Agent checks which models and services the project needs. If a
+required credential is missing, it will ask you for it and explain what it is used for. You have two
+options:
+
+1. **Use Hypit's recommended Hypit.ai OAuth login.** Tell the Agent to log in through hypit.ai. This
+ single OAuth flow covers all models that Hypit provides through its hosted service, so you do not
+ need to collect separate keys for each model.
+2. **Use your own provider keys.** You can tell the Agent which models to use and provide API keys
+ from the corresponding providers. The Agent will request only the keys needed for this project
+ and will not ask for unrelated credentials.
+
+Never paste a secret into a public document or commit it to Git. The Agent stores credentials using
+the configured secure credential store.
+
+## 4. Let the Agent do the production work
After you approve the brief, the Agent works through the project without requiring you to write
source code. The exact sequence depends on the video, but it generally includes:
@@ -113,7 +116,7 @@ This is why the workflow can feel like directing a small production team: the Ag
cuts, Gemini analysis, package discovery, source authoring, mock comparison and repairs. You are
not expected to open an editor, write a component, or run a sequence of build commands yourself.
-## 4. Review the mock Studio
+## 5. Review the mock Studio
When the first pass is ready, the Agent opens a Studio mock for you. A mock is a review version: it
uses deterministic stand-ins for media that has not been generated yet, while preserving the real
@@ -132,7 +135,7 @@ while she is speaking. Please fix those issues and show me the mock again.
The Agent edits the source, reruns the relevant checks and returns an updated mock. You review the
result again; you do not need to locate the corresponding SVML line.
-## 5. Approve and submit the paid Build
+## 6. Approve and submit the paid Build
Only after you explicitly approve the mock should you ask for the paid Build:
@@ -145,7 +148,7 @@ It then submits the Build, follows its progress and reports any provider or runt
language. A paid Build is the step that performs the real generation, media processing and final
rendering; the earlier mock does not silently trigger those billable operations.
-## 6. Review the paid result
+## 7. Review the paid result
When the Build finishes, the Agent opens the resulting video and provides the saved output. Watch
the final result, not only the Studio mock. Real generated shots can differ from their mock
@@ -153,16 +156,15 @@ stand-ins, so check the generated host, B-roll, audio, captions, transitions, fr
quality. If the result needs a correction, tell the Agent what to change. It will revise the source
and guide you through another review rather than asking you to edit the rendered MP4 by hand.
-## 7. Request natural-language changes
+## 8. Request natural-language changes
After a successful Build, you can continue directing the project conversationally. You may change
the host, replace B-roll, alter the ranking-board treatment, translate the script, adjust the tone or
switch the aspect ratio. For example:
```text
-Keep the script and ranking order, but replace the host with a calm male presenter. Use hand-held
-street footage for the B-roll, make the board look like a paper sports magazine, and keep the captions
-large enough for a 9:16 phone screen.
+Keep the script and ranking order, but replace the host with a calm male presenter and make the board
+look like a paper sports magazine.
```
The Agent traces each request to the relevant source and component, preserves what you asked it to
@@ -170,7 +172,7 @@ keep, and shows a new mock before another paid Build. You can also ask it to use
service or Runtime Profile when that is supported by your project; the Agent explains any new
credential or cost requirement first.
-## 8. Create multiple variants in parallel
+## 9. Create multiple variants in parallel
When you want several versions, first tell the Agent the dimensions that may vary. It will discuss
the intended directions with you, for example:
@@ -185,11 +187,3 @@ directions, it creates a separate project scope for each variant and dispatches
in parallel. Each child keeps the agreed boundaries, checks its own layout and source, and reports
back to the main Agent. You receive a clear list of completed variants and any decision that still
needs your input; you do not have to coordinate the child Agents yourself.
-
-## If you want to write SVML or SVS yourself
-
-The Agent workflow above is the recommended first experience. The next page begins the manual
-authoring path: [Script](./quickstart/script.md) explains how to write the first Author Source,
-connect semantic segments and continue into SVS Recipes, media, timing, Tracks and rendering. You
-can switch between the two workflows at any time because the Agent produces ordinary, editable Hypit
-source files.
diff --git a/docs/zh/quickstart.md b/docs/zh/quickstart.md
index 6d16fcf7..b4d6d26e 100644
--- a/docs/zh/quickstart.md
+++ b/docs/zh/quickstart.md
@@ -18,30 +18,20 @@ Coding Agent 描述想要的结果;Agent 会通过 `/hypit` skill 准备项目
- 一个能够使用 skill 的 Coding Agent,例如 Claude Code 或 Codex;
- 一个允许 Agent 创建视频项目的文件夹;
- 一条用于复刻的视频,或一份用于原创视频的创意简报;
-- 你所选择的模型和服务的凭据。Agent 只会在当前 Runtime Profile 确实需要时索取这些凭据。
-你不需要克隆 Hypit 仓库。Agent 可以替你安装可复用的 `/hypit` skill 和 Hypit Distribution。需要执行
-安装命令时,Agent 会先解释命令的作用,再在项目目录中完成操作。
+你不需要克隆 Hypit 仓库。先按第一步的命令安装可复用的 `/hypit` skill;之后 Agent 可以替你准备 Hypit
+Distribution。需要执行其他安装命令时,Agent 会先解释命令的作用,再在项目目录中完成操作。
-## 第一步:使用 Hypit skill 并提供凭据
+## 第一步:安装 Hypit skill
-在准备存放项目的文件夹中打开 Coding Agent,然后要求它使用 Hypit 并登录:
+在准备存放项目的文件夹中打开 Coding Agent,然后先安装 Hypit skill:
-```text
-/hypit login
-```
-
-skill 会引导你完成登录,并使用已配置的凭据存储安全保存登录信息。你也可以明确要求 Agent 使用自己的
-Provider key,例如:“使用我自己的 KIE 和 Gemini key,不要创建新的账号。”Agent 会判断当前视频真正
-需要哪些 key,说明每个 key 的用途,并避免索取无关凭据。不要把密钥粘贴到公开文档中,也不要提交到 Git。
-
-如果尚未安装 skill,可以直接告诉 Agent:
-
-```text
-安装 Hypit skill,然后帮我登录并检查环境是否就绪。
+```bash
+npx skills add hypit-ai/hypit -g
```
-Agent 会完成安装和本地环境检查。你不需要记忆包管理器、Runtime Profile 或各个服务的专属配置步骤。
+`-g` 选项会让 skill 对任何项目文件夹中的新 Agent 会话都可用。你不需要克隆 Hypit 仓库,也不需要
+理解这条命令安装了哪些文件。skill 可用后,在视频项目文件夹中开启新的 Agent 对话,然后继续下面的步骤。
## 第二步:描述你想制作的视频
@@ -53,14 +43,13 @@ Agent 会完成安装和本地环境检查。你不需要记忆包管理器、Ru
会自动把路径填入消息。也可以手动写出路径:
```text
-/hypit Clone this video and preserve its pacing, shot structure, captions, B-roll rhythm and sound design:
-/Users/me/Desktop/reference.mp4
+/hypit 克隆这条视频:/path/to/video.mp4
```
如果需要改变内容,也可以一并说明,例如更换主持人、语言、产品、画幅比例、视觉风格或行动号召。参考视频
会被当作剪辑结构的证据;Agent 不会只给你一份分析报告,而是会创建一套可以审阅、修改和 Build 的完整视频程序。
-### 根据创意简报原创一条视频
+### 原创一条视频
你可以像给一位真人制片人写 brief 一样描述想法。例如:
@@ -74,7 +63,19 @@ Arcads、Higgsfield、Seedance 和 CapCut,对每个竞品给出公平而简洁
要求由哪个 package、模型、字幕系统或动画原语实现。Agent 会把 brief 转换成完整计划;只有在某个创意
决定无法安全推断时,才会提出简短的补充问题。
-## 第三步:让 Agent 完成制作
+## 第三步:按 Agent 提示提供凭据
+
+在你描述视频之后,Agent 会检查项目需要哪些模型和服务。如果缺少必要凭据,Agent 会向你询问,并说明
+该凭据的用途。你可以选择两种方式:
+
+1. **使用 Hypit 推荐的 hypit.ai OAuth 登录。** 告诉 Agent 通过 hypit.ai 登录。一次 OAuth 登录即可覆盖
+ Hypit 托管服务提供的所有模型,不需要为每个模型分别收集 key。
+2. **使用你自己的 Provider key。** 你可以直接和 Agent 约定要使用哪些模型,并提供相应 Provider 的 API key。
+ Agent 只会索取当前项目需要的 key,不会询问无关凭据。
+
+不要把密钥粘贴到公开文档中,也不要提交到 Git。Agent 会使用已配置的安全凭据存储来保存这些凭据。
+
+## 第四步:让 Agent 完成制作
你确认 brief 后,Agent 会在不要求你编写源码的情况下推进项目。具体顺序会因视频而不同,但通常包括:
@@ -88,7 +89,7 @@ Arcads、Higgsfield、Seedance 和 CapCut,对每个竞品给出公平而简洁
因此,这个流程更像是在指导一支小型制作团队:切分 shot、Gemini 分析、提出窄问题、查找 package、写源码、生成 mock、进行比对并修复,全部由 Agent 完成。你不需要打开编辑器、编写组件,也不需要自己执行一连串 Build 命令。
-## 第四步:查看 mock Studio
+## 第五步:查看 mock Studio
第一版完成后,Agent 会为你打开 Studio mock。Mock 是审阅版本:尚未生成的媒体会使用确定性的临时素材代替,但真实的时序、布局、字幕、转场和 Track 关系都会保留。它要回答的是“这支视频是否成立”,而不是在你确认之前悄悄执行付费生成或最终渲染。
@@ -100,7 +101,7 @@ ranking 板出现得太晚,手机上字幕太小,而且主持人说话时被
Agent 会修改源码,重新执行相关检查,并返回更新后的 mock。你只需再次审阅,不需要自己寻找对应的 SVML 行。
-## 第五步:确认后提交付费 Build
+## 第六步:确认后提交付费 Build
只有在你明确确认 mock 后,才要求 Agent 提交付费 Build:
@@ -110,21 +111,21 @@ mock 没有问题。请提交付费 Build 并生成最终视频。
开始之前,Agent 会向你总结将使用的模型、预计执行的外部工作和预估费用。随后它会提交 Build、跟踪进度,并用书面化的语言报告 Provider 或 Runtime 问题。付费 Build 才会执行真实的素材生成、媒体处理和最终渲染;之前的 mock 不会自动触发这些计费操作。
-## 第六步:查看付费 Build 的结果
+## 第七步:查看付费 Build 的结果
Build 完成后,Agent 会打开最终视频并提供保存后的输出文件。请查看最终结果,而不只是 Studio mock。真实生成的主持人、B-roll 和音频可能与 mock 的临时素材不同,因此要检查画面、字幕、转场、构图、声音和导出质量。如果需要修改,直接告诉 Agent 要改什么。Agent 会修改源码并带你重新审阅,而不是让你手动编辑已经渲染出的 MP4。
-## 第七步:继续用自然语言提出修改
+## 第八步:继续用自然语言提出修改
Build 成功后,你仍然可以用对话继续指导项目。你可以更换主持人、替换 B-roll、改变 ranking 板的样式、翻译台词、调整语气或切换画幅比例。例如:
```text
-保留台词和排名顺序,但把主持人换成沉稳的男主持。B-roll 改用手持街头素材,ranking 板做成纸质体育杂志风格,并确保 9:16 手机画面上的字幕足够大。
+保留台词和排名顺序,但把主持人换成沉稳的男主持,并把 ranking 板做成纸质体育杂志风格。
```
Agent 会把每项要求追踪到对应的源码和组件,保留你要求不变的部分,并在再次付费 Build 前先给你看新的 mock。如果项目支持,你也可以要求切换到另一种主持人服务或 Runtime Profile;Agent 会先说明新增的凭据和费用要求。
-## 第八步:并行制作多个变体
+## 第九步:并行制作多个变体
如果需要多个版本,先告诉 Agent 哪些维度可以变化。它会与你讨论具体方向,例如:
@@ -134,7 +135,3 @@ Agent 会把每项要求追踪到对应的源码和组件,保留你要求不
- 加快剪辑、增加 B-roll 密度的版本。
Agent 会先确认所有版本共用的内容,以及每个版本允许改变的范围。你确认方向后,它会为每个变体创建独立的项目范围,并下发子 Agent 并行制作。每个子 Agent 都会遵守已确认的边界,检查自己的布局和源码,再向主 Agent 汇报。最终你会得到清晰的变体完成列表,以及仍需要你决定的事项;不需要你亲自协调这些子 Agent。
-
-## 如果你想亲自编写 SVML 或 SVS
-
-上面的 Agent 工作流适合作为第一次体验。下一页开始介绍手动编写路径:[Script](./quickstart/script.md) 会说明如何编写第一份 Author Source、连接语义 Segment,并继续学习 SVS Recipe、媒体、时序、Track 和渲染。两种工作流可以随时切换,因为 Agent 生成的仍然是普通、可编辑的 Hypit 源文件。