anycap-media-production

Compare original and translation side by side

🇺🇸

Original

English

🇨🇳

Translation

Chinese

AnyCap Media Production

AnyCap媒体生产

Read this entire file before starting. It covers the full production workflow across image, video, music, and audio -- including iterative refinement with human feedback.

Workflow guide for producing media assets with AnyCap. Covers image, video, music, and audio -- from initial generation through iterative refinement to delivery.

This skill is about how to produce media. For CLI command reference and parameters, read the

anycap-cli

skill.

开始前请通读整个文档。 本文档涵盖了图像、视频、音乐和音频的完整生产工作流——包括结合人工反馈的迭代优化流程。

使用AnyCap制作媒体资产的工作流指南。涵盖图像、视频、音乐和音频——从初始生成到迭代优化再到交付的全流程。

本技能介绍如何制作媒体资产。如需CLI命令参考和参数说明，请查阅

anycap-cli

技能文档。

Prerequisites

前提条件

AnyCap CLI must be installed and authenticated. Read the

anycap-cli

skill if setup is needed.

必须安装并完成AnyCap CLI的身份验证。如需进行设置，请查阅

anycap-cli

技能文档。

Quick Reference

快速参考

Media	Generate	Refine	Typical duration
Image	`anycap image generate`	Annotate + image-to-image	5-30s
Video	`anycap video generate`	Re-generate with adjusted params	30-120s
Music	`anycap music generate`	Re-generate with adjusted prompt	30-90s
Audio	`anycap audio generate`	Re-generate with adjusted prompt or references	Model-dependent

All generation commands follow the same pattern:

1. Discover models    anycap {cap} models
2. Check schema       anycap {cap} models <model> schema [--mode <mode>]
3. Generate           anycap {cap} generate --model <model> --prompt "..." -o output.ext

Always choose model IDs from the live model catalog and inspect the schema for the selected mode before relying on model-specific parameters. Always use

-o

with a descriptive filename.

媒体类型	生成命令	优化方式	典型耗时
图像	`anycap image generate`	标注 + 图生图	5-30秒
视频	`anycap video generate`	调整参数重新生成	30-120秒
音乐	`anycap music generate`	调整提示词重新生成	30-90秒
音频	`anycap audio generate`	调整提示词或参考内容重新生成	取决于模型

所有生成命令遵循相同模式：

1. 发现模型    anycap {cap} models
2. 查看 schema       anycap {cap} models <model> schema [--mode <mode>]
3. 生成内容           anycap {cap} generate --model <model> --prompt "..." -o output.ext

请始终从实时模型目录中选择模型ID，并在使用模型特定参数前查看所选模式的schema。请始终使用

-o

参数指定具有描述性的文件名。

Image Production

图像制作

Text-to-Image

文本生成图像

Generate an image from a text prompt:

bash

anycap image generate \
  --prompt "a cozy home office with a wooden desk, laptop, coffee cup, and plants by the window" \
  --model <model-id> \
  -o workspace-v1.png

通过文本提示词生成图像：

bash

anycap image generate \
  --prompt "a cozy home office with a wooden desk, laptop, coffee cup, and plants by the window" \
  --model <model-id> \
  -o workspace-v1.png

Image-to-Image (Edit / Transform)

图生图（编辑/转换）

Use

--mode image-to-image

with a reference image to edit or transform an existing image:

bash

anycap image generate \
  --prompt "make it a watercolor painting" \
  --model <model-id> \
  --mode image-to-image \
  --param images=./photo.png \
  -o photo-watercolor.png

Reference images can be local paths or URLs. The CLI handles upload automatically.

使用

--mode image-to-image

模式并传入参考图像，即可编辑或转换现有图像：

bash

anycap image generate \
  --prompt "make it a watercolor painting" \
  --model <model-id> \
  --mode image-to-image \
  --param images=./photo.png \
  -o photo-watercolor.png

参考图像可以是本地路径或URL。CLI会自动处理上传操作。

Multiple Reference Images

多参考图像

Some models accept multiple reference images for style transfer, composition blending, or subject-driven generation. Use JSON array syntax to pass multiple files:

bash

undefined

部分模型支持传入多张参考图像，用于风格迁移、构图融合或基于主体的生成。使用JSON数组语法传入多个文件：

bash

undefined

Combine style from one image with composition from another

将一张图像的风格与另一张的调色板融合

anycap image generate
--prompt "merge the architectural style of the first image with the color palette of the second"
--model <model-id>
--mode image-to-image
--param images='["./style-ref.png","./color-ref.png"]'
-o blended.png

Mix local files and URLs

混合本地文件和URL

anycap image generate
--prompt "a portrait in the style of the reference images"
--model <model-id>
--mode image-to-image
--param images='["./local-ref.png","https://example.com/style-ref.jpg"]'
-o portrait-styled.png


Tips:
- Use JSON array syntax `'["path1","path2"]'` -- repeating `--param images=` overwrites rather than appends.
- Local file paths inside the array are auto-uploaded, same as single-file mode.
- Not all models support multiple references. Check the model schema first. When unsupported, the model typically uses only the first image.


提示：
- 使用JSON数组语法`'["path1","path2"]'`——重复使用`--param images=`会覆盖之前的设置，而非追加。
- 数组中的本地文件路径会自动上传，与单文件模式相同。
- 并非所有模型都支持多参考图像。请先查看模型的schema。如果不支持，模型通常只会使用第一张图像。

Iterative Refinement with Annotation

基于标注的迭代优化

When text prompts alone cannot describe the desired edit precisely ("move this", "remove that specific thing", "change the color of this area"), use the annotation workflow. For the full annotation guide -- including URL/video review, headless access, recording analysis, and multi-user collaboration -- read the

anycap-human-interaction

skill.

mermaid

graph TD
    A[Start: concept or existing image] --> B{Have an image?}
    B -->|No| C[Generate initial image]
    B -->|Yes| D[Human annotates the image]
    C --> D
    D --> E[Build prompt from annotations]
    E --> F[Generate with image-to-image]
    F --> G[Show result to human]
    G --> H{Satisfied?}
    H -->|Yes| I[Done -- deliver final asset]
    H -->|No| D

当仅靠文本提示词无法精准描述所需编辑（如“移动这个元素”、“移除那个特定物体”、“更改这个区域的颜色”）时，请使用标注工作流。如需完整的标注指南——包括URL/视频审核、无头访问、录制分析和多用户协作，请查阅

anycap-human-interaction

技能文档。

mermaid

graph TD
    A[开始：概念或现有图像] --> B{已有图像？}
    B -->|否| C[生成初始图像]
    B -->|是| D[人工标注图像]
    C --> D
    D --> E[基于标注构建提示词]
    E --> F[通过图生图生成新图像]
    F --> G[向人工展示结果]
    G --> H{满意？}
    H -->|是| I[完成——交付最终资产]
    H -->|否| D

Step 1: Generate or Use an Existing Image

步骤1：生成或使用现有图像

bash

anycap image generate \
  --prompt "a landing page hero banner with mountains and sunrise" \
  --model <model-id> \
  -o banner-v1.png

bash

anycap image generate \
  --prompt "a landing page hero banner with mountains and sunrise" \
  --model <model-id> \
  -o banner-v1.png

Step 2: Annotate

步骤2：标注

Open the annotation tool so the human can visually mark regions, describe desired changes, and optionally record a narrated walkthrough. Multiple users can collaborate on the same session in real-time.

For agent workflows (non-blocking, recommended):

bash

anycap annotate banner-v1.png --no-wait -o banner-v1-annotated.png

打开标注工具，让人工可视化标记区域、描述所需更改，还可选择录制旁白讲解。多名用户可实时协作处理同一会话。

适用于Agent工作流（非阻塞，推荐使用）：

bash

anycap annotate banner-v1.png --no-wait -o banner-v1-annotated.png

Returns: {session, url, poll_command, stop_command}

返回结果：{session, url, poll_command, stop_command}


Show the URL to the human and ask them to annotate. Multiple people can open the same URL to collaborate. Wait for the human to confirm they are done, then:

```bash


将URL展示给人工并请求其进行标注。多人可打开同一URL协作。等待人工确认完成后，执行以下命令：

```bash

Fetch the result (single call, no loop)

获取结果（单次调用，无需循环）

anycap annotate poll --session <session_id>

If recording exists, analyze it for visual understanding

如果存在录制文件，分析其以理解视觉需求

anycap actions video-read --file .anycap/annotate/<session_id>/recording.webm
--instruction "Describe what changes the user wants"

Clean up

清理会话

anycap annotate stop --session <session_id>


**For interactive sessions** (human is at the terminal):

```bash
anycap annotate banner-v1.png -o banner-v1-annotated.png

anycap annotate stop --session <session_id>


**适用于交互式会话**（人工在终端操作）：

```bash
anycap annotate banner-v1.png -o banner-v1-annotated.png

Blocks until Done click, outputs annotation JSON

会阻塞直到点击“完成”，输出标注JSON


The annotation tool supports four tools: Rectangle (`R`), Arrow (`A`), Point (`P`), Freehand (`F`). Each annotation gets a numbered marker and a text label.


标注工具支持四种工具：矩形（`R`）、箭头（`A`）、点（`P`）、自由手绘（`F`）。每个标注都会获得一个编号标记和文本标签。

Step 3: Build a Prompt from Annotations

步骤3：基于标注构建提示词

The annotation output contains structured data. Translate each label into a coherent prompt:

json

{
  "annotations": [
    {"id": 1, "type": "rect", "label": "Replace with a standing desk"},
    {"id": 2, "type": "point", "label": "Add a cat sitting here"},
    {"id": 3, "type": "freehand", "label": "This area should be a bookshelf"}
  ]
}

Prompt: "#1: Replace the desk with a standing desk. #2: Add a cat sitting at the marked position. #3: Transform the outlined area into a bookshelf. Keep all other elements unchanged."

Rules:

Reference each annotation by its number (#1, #2, etc.)
Include the human's exact label text
Add "Keep all other elements unchanged" to preserve unmodified areas

标注输出包含结构化数据。将每个标签转换为连贯的提示词：

json

{
  "annotations": [
    {"id": 1, "type": "rect", "label": "Replace with a standing desk"},
    {"id": 2, "type": "point", "label": "Add a cat sitting here"},
    {"id": 3, "type": "freehand", "label": "This area should be a bookshelf"}
  ]
}

提示词："#1: 将书桌替换为站立式书桌。#2: 在标记位置添加一只猫。#3: 将勾勒的区域改为书架。保持所有其他元素不变。"

规则：

通过编号（#1、#2等）引用每个标注
包含人工提供的准确标签文本
添加“保持所有其他元素不变”以保留未修改区域

Step 4: Apply the Edit

步骤4：应用编辑

Use the annotated image (with visual markers) as the reference:

bash

anycap image generate \
  --prompt "#1: Replace the desk with a standing desk. #2: Add a cat. Keep all other elements unchanged." \
  --model <model-id> \
  --mode image-to-image \
  --param images=./banner-v1-annotated.png \
  -o banner-v2.png

使用带标注的图像（含视觉标记）作为参考：

bash

anycap image generate \
  --prompt "#1: Replace the desk with a standing desk. #2: Add a cat. Keep all other elements unchanged." \
  --model <model-id> \
  --mode image-to-image \
  --param images=./banner-v1-annotated.png \
  -o banner-v2.png

Step 5: Iterate

步骤5：迭代

If the human wants more changes, use the latest version as input and repeat from Step 2. Version filenames (

v1

v2

v3

) so the human can compare and revert.

如果人工需要更多更改，使用最新版本作为输入，从步骤2开始重复操作。为文件名添加版本号（

v1

、

v2

、

v3

），以便人工对比和回退。

Image Tips

图像制作提示

Start broad, refine narrow. First generation nails the composition. Annotation iterations handle targeted adjustments.
One thing at a time. If multi-region edits produce poor results, try one annotation per pass.
Annotated image only. Pass only the annotated image as the reference. Most models understand numbered markers and remove them from the output.

先宽泛生成，再精准优化。首次生成确定构图，标注迭代处理针对性调整。
一次处理一个元素。如果多区域编辑效果不佳，尝试每次仅处理一个标注。
仅传入带标注的图像。仅将带标注的图像作为参考。大多数模型能识别编号标记并在输出中移除它们。

Video Production

视频制作

Text-to-Video

文本生成视频

bash

anycap video generate \
  --prompt "a cat walking on the beach at sunset, cinematic, slow motion" \
  --model <model-id> \
  -o cat-beach.mp4

bash

anycap video generate \
  --prompt "a cat walking on the beach at sunset, cinematic, slow motion" \
  --model <model-id> \
  -o cat-beach.mp4

Image-to-Video

图像生成视频

Animate a still image:

bash

anycap video generate \
  --prompt "gentle camera pan across the landscape, wind blowing through trees" \
  --model <model-id> \
  --mode image-to-video \
  --param images=./landscape.png \
  -o landscape-animated.mp4

This is powerful for combining with image generation: generate a still image first, then animate it.

将静态图像动画化：

bash

anycap video generate \
  --prompt "gentle camera pan across the landscape, wind blowing through trees" \
  --model <model-id> \
  --mode image-to-video \
  --param images=./landscape.png \
  -o landscape-animated.mp4

这一功能与图像生成结合使用时非常强大：先生成静态图像，再将其动画化。

Video Production Workflow

视频制作工作流

mermaid

graph LR
    A[Text prompt] --> B[Generate image]
    B --> C{Animate?}
    C -->|Yes| D[image-to-video]
    C -->|No| E[Done]
    A --> F[text-to-video]
    F --> E
    D --> E

For best results with image-to-video:

Generate a high-quality still image first (iterate with annotation if needed)
Use the final image as the reference for video generation
Keep the video prompt focused on motion and camera movement, not scene description

mermaid

graph LR
    A[文本提示词] --> B[生成图像]
    B --> C{是否动画化？}
    C -->|是| D[图像生成视频]
    C -->|否| E[完成]
    A --> F[文本生成视频]
    F --> E
    D --> E

为获得最佳的图像生成视频效果：

先生成高质量的静态图像（如有需要，通过标注迭代优化）
使用最终图像作为视频生成的参考
视频提示词专注于运动和镜头移动，而非场景描述

Video Tips

视频制作提示

Video generation takes 30-120s. Use async execution when your runtime supports it.
Check model schema for supported parameters (
```
aspect_ratio
```
,
```
duration
```
, etc.).
Different models excel at different styles. Check available models with
```
anycap video models
```
.

视频生成耗时30-120秒。如果运行时支持，请使用异步执行。
查看模型schema以了解支持的参数（
```
aspect_ratio
```
、
```
duration
```
等）。
不同模型擅长不同风格。使用
```
anycap video models
```
查看可用模型。

Music Production

音乐制作

Text-to-Music

文本生成音乐

bash

anycap music generate \
  --prompt "upbeat electronic track with synth leads and driving bass, 120 BPM" \
  --model <model-id> \
  -o background-track.mp3

Music generation may return multiple clips. Extract the first:

bash

anycap music generate --prompt "..." --model <model-id> -o track.mp3 \
  | jq -r '.outputs[0].local_path'

bash

anycap music generate \
  --prompt "upbeat electronic track with synth leads and driving bass, 120 BPM" \
  --model <model-id> \
  -o background-track.mp3

音乐生成可能返回多个片段。提取第一个片段：

bash

anycap music generate --prompt "..." --model <model-id> -o track.mp3 \
  | jq -r '.outputs[0].local_path'

Music Tips

音乐制作提示

Be specific about genre, tempo, instruments, and mood in prompts.
Music generation takes 30-90s. Use async execution when possible.
Check model parameters via schema -- some models support
```
duration
```
,
```
genre
```
,
```
tags
```
.

在提示词中明确说明流派、节奏、乐器和情绪。
音乐生成耗时30-90秒。如果可能，请使用异步执行。
通过schema查看模型参数——部分模型支持
```
duration
```
、
```
genre
```
、
```
tags
```
。

Audio Production

音频制作

Audio generation covers speech synthesis, dialogue, and complete audio scenes generated from text or reference media. A scene can combine voices, background music, ambience, and sound effects; use the separate music capability when the deliverable is primarily a song or instrumental track.

bash

undefined

音频生成涵盖语音合成、对话以及通过文本或参考媒体生成完整音频场景。一个场景可包含人声、背景音乐、环境音和音效；当交付内容主要是歌曲或器乐时，请使用单独的音乐生成功能。

bash

undefined

Discover live modes and controls first

先查看可用模式和控制选项

anycap audio models <model-id> schema --mode text-to-audio

Generate speech with supporting ambience from text

通过文本生成带环境音的语音

anycap audio generate
--prompt 'A calm narrator says: "Welcome to the evening program." Soft room ambience underneath.'
--model <model-id>
--mode text-to-audio
-o evening-introduction.mp3

Guide a new voice performance with reference audio

参考音频指导新的语音生成

anycap audio generate
--prompt "create a new spoken welcome with the reference delivery style"
--model <model-id>
--mode audio-to-audio
--param audios=./reference.wav
-o guided-welcome.mp3

Create a narrated audio scene from an image

通过图像生成带旁白的音频场景

anycap audio generate
--prompt "a guide describes this scene while matching ambience plays underneath"
--model <model-id>
--mode image-to-audio
--param images=./scene.png
-o narrated-scene.mp3


Use the live schema for reference limits and audio controls. Local reference files are uploaded automatically. The JSON output preserves duration, size, subtitle, and usage data when the provider returns them.

anycap audio generate
--prompt "a guide describes this scene while matching ambience plays underneath"
--model <model-id>
--mode image-to-audio
--param images=./scene.png
-o narrated-scene.mp3


查看实时schema以了解参考限制和音频控制选项。本地参考文件会自动上传。当提供商返回相关数据时，JSON输出会保留时长、大小、字幕和使用数据。

Delivery

交付

When the asset is ready, deliver using the appropriate method:

bash

undefined

当资产准备就绪后，使用合适的方式交付：

bash

undefined

Share via Drive (generates a shareable link)

通过云端硬盘分享（生成可分享链接）

anycap drive upload banner-final.png anycap drive share banner-final.png

Publish as a web page

发布为网页

anycap page deploy ./site-directory

undefined

anycap page deploy ./site-directory

undefined

Multi-Media Production Example

多媒体制作示例

A complete workflow producing a promotional package:

bash

undefined

制作宣传素材的完整工作流：

bash

undefined

1. Generate hero image

1. 生成主视觉图像

anycap image generate
--prompt "modern SaaS dashboard with data visualizations, dark mode, purple accents"
--model <image-model-id> -o hero-v1.png

2. Refine via annotation (agent asks human to mark changes)

2. 通过标注优化（Agent请求人工标记更改）

anycap annotate hero-v1.png --no-wait -o hero-v1-annotated.png

... human annotates, agent polls result ...

... 人工标注，Agent获取结果 ...

anycap image generate
--prompt "#1: Make the chart larger. #2: Change accent color to blue."
--model <image-model-id> --mode image-to-image
--param images=./hero-v1-annotated.png -o hero-v2.png

3. Create an animated version

3. 创建动画版本

anycap video generate
--prompt "slow zoom into the dashboard, data points animate in sequentially"
--model <video-model-id> --mode image-to-video
--param images=./hero-v2.png -o hero-animation.mp4

4. Generate background music

4. 生成背景音乐

anycap music generate
--prompt "ambient tech background music, minimal, clean, 90 BPM"
--model <music-model-id> -o background-music.mp3

5. Deliver

5. 交付

anycap drive upload hero-v2.png hero-animation.mp4 background-music.mp3

undefined

anycap drive upload hero-v2.png hero-animation.mp4 background-music.mp3

undefined