手办化?coser化?如何让你的机器人也拥有同样的功能?

看到别人的bot可以手办化?你也想玩xx化?


本篇将教你如何使用Chatluna实现这样的功能!

从入门Chatluna开始!

前排提醒!

此功能需要你配置可以绘图的模型API和key
相关API和key,可能会需要涉及付费!

价格因你所使用的厂商、模型而有所区别。(当然如果你能找到免费绘图模型的那这个功能就是免费的((((()

注意使用的模型是gemini 2.5 image preview,确保你的API可以使用此模型!

如果你希望使用openai like的编辑图片接口,请使用另一个插件: https://www.npmjs.com/package/koishi-plugin-lmarena-api

1 个赞

首先你需要前往插件市场安装对应的Chatluna插件~

需要包含以下三个插件哦

chatluna主插件

chatluna 

chatluna模型适配器(这里以openai类的作为教程演示)

chatluna-openai-like-adapter 

chatluna快捷指令插件

chatluna-actions 

如果你遇到了插件市场访问问题

请前往 为什么我找不到downloads插件 - #4,来自 shangxue 解决后回到本帖~


1 个赞

第一步,开启chatluna主插件

1 个赞

第二步 配置openai适配器

前往你的模型服务商,找到对应的请求地址和APIkey

注意这里,教程中演示的是来自 https://api.bailili.top/ 中转站的模型服务。

如果你从未使用过中转站服务,那么推荐你来这里试试~

优点:国内可访问、较官网更低价、模型更多、并发优秀…


如何使用 api.bailili.top ?

如何使用 api.bailili.top
  1. 打开 New API

注册账号,这里推荐使用QQ邮箱注册。

  1. 复制你的秘钥

  1. 如上述步骤配置 chatluna-openai-like-adapter 插件

3.1 右侧填入请求地址

https://api.bailili.top/v1

3.2 左侧填入你的秘钥

(注意不要把下面这个模拟秘钥复制进去哦)

sk-hy...................................7T2

开启 chatluna-openai-like-adapter 插件

验证服务可用 / 选择模型

回到chatluna主插件

在配置项较下方的位置

选择你希望使用的默认模型

如果此处你无法选择模型,可以尝试重启koishi

如果重启之后仍然不能选择模型,则说明你的 chatluna-openai-like-adapter 插件没有开启,或者提供的服务不可用。
可以考虑更换服务商等方法…

1 个赞

第三步 配置 chatluna-actions 插件

此步是实现 手办化 功能的重要步骤


来到 chatluna-actions 插件配置项

  1. 添加自定义指令

  1. 快速导入 手办化 指令配置
{
  "command": "手办化",
  "model": "openai-like/gemini-2.5-flash-image-preview",
  "description": "把输入图片变成手办",
  "preset": "chatgpt",
  "inputPrompt": "Your task is to create a photorealistic, masterpiece-quality image of a 1/7 scale commercialized figurine based on the user\\'s character. The final image must be in a realistic style and environment.\n\n**Crucial Instruction on Face & Likeness:** The figurine\\'s face is the most critical element. It must be a perfect, high-fidelity 3D translation of the character from the source image. The sculpt must be sharp, clean, and intricately detailed, accurately capturing the original artwork\\'s facial structure, eye style, expression, and hair. The final result must be immediately recognizable as the same character, elevated to a premium physical product standard. Do NOT generate a generic or abstract face.\n\n**Scene Composition (Strictly follow these details):**\n1. **Figurine & Base:** Place the figure on a computer desk. It must stand on a simple, circular, transparent acrylic base WITHOUT any text or markings.\n2. **Computer Monitor:** In the background, a computer monitor must display 3D modeling software (like ZBrush or Blender) with the digital sculpt of the very same figurine visible on the screen.\n3. **Artwork Display:** Next to the computer screen, include a transparent acrylic board with a wooden base. This board holds a print of the original 2D artwork that the figurine is based on.\n4. **Environment:** The overall setting is a desk, with elements like a keyboard to enhance realism. The lighting should be natural and well-lit, as if in a room.',"
}

  1. 确保模型支持绘图!

这里选择的是 gemini-2.5-flash-image-preview


第四步 开启插件


第五步 验证功能

在支持图文功能的平台,验证你的机器人功能!

注意!手办化 文字后面还要加一个空格,然后再加一个图片,也就是

手办化 [图片]

而不是这样哦:

手办化[图片]

2 个赞

更多指令?

如何导入更多指令?

如何导入更多指令?————点我查看教程


下面是更多的指令配置分享

如果大家有很好玩的提示词,也欢迎分享!

{
    "command": "手办化",
    "model": "openai-like/gemini-2.5-flash-image-preview",
    "description": "把输入图片变成手办",
    "preset": "chatgpt",
    "inputPrompt": "Your task is to create a photorealistic, masterpiece-quality image of a 1/7 scale commercialized figurine based on the user\\'s character. The final image must be in a realistic style and environment.\n\n**Crucial Instruction on Face & Likeness:** The figurine\\'s face is the most critical element. It must be a perfect, high-fidelity 3D translation of the character from the source image. The sculpt must be sharp, clean, and intricately detailed, accurately capturing the original artwork\\'s facial structure, eye style, expression, and hair. The final result must be immediately recognizable as the same character, elevated to a premium physical product standard. Do NOT generate a generic or abstract face.\n\n**Scene Composition (Strictly follow these details):**\n1. **Figurine & Base:** Place the figure on a computer desk. It must stand on a simple, circular, transparent acrylic base WITHOUT any text or markings.\n2. **Computer Monitor:** In the background, a computer monitor must display 3D modeling software (like ZBrush or Blender) with the digital sculpt of the very same figurine visible on the screen.\n3. **Artwork Display:** Next to the computer screen, include a transparent acrylic board with a wooden base. This board holds a print of the original 2D artwork that the figurine is based on.\n4. **Environment:** The overall setting is a desk, with elements like a keyboard to enhance realism. The lighting should be natural and well-lit, as if in a room.'"
}
{
    "command": "手办化2",
    "model": "openai-like/gemini-2.5-flash-image-preview",
    "description": "把输入图片变成手办,并展示建模过程和包装盒",
    "preset": "chatgpt",
    "inputPrompt": "Use the nano-banana model to create a 1/7 scale commercialized figure of thecharacter in the illustration, in a realistic styie and environment.Place the figure on a computer desk, using a circular transparent acrylic basewithout any text.On the computer screen, display the ZBrush modeling process of the figure.Next to the computer screen, place a BANDAl-style toy packaging box printedwith the original artwork."
}
{
    "command": "手办化3",
    "model": "openai-like/gemini-2.5-flash-image-preview",
    "description": "把输入图片变成手办,并展示在商业包装中",
    "preset": "chatgpt",
    "inputPrompt": "Your primary mission is to accurately convert the subject from the user\\'s photo into a photorealistic, masterpiece quality, 1/7 scale PVC figurine, presented in its commercial packaging.\n\n**Crucial First Step: Analyze the image to identify the subject\\'s key attributes (e.g., human male, human female, animal, specific creature) and defining features (hair style, clothing, expression). The generated figurine must strictly adhere to these identified attributes.** This is a mandatory instruction to avoid generating a generic female figure.\n\n**Top Priority - Character Likeness:** The figurine\\'s face MUST maintain a strong likeness to the original character. Your task is to translate the 2D facial features into a 3D sculpt, preserving the identity, expression, and core characteristics. If the source is blurry, interpret the features to create a sharp, well-defined version that is clearly recognizable as the same character.\n\n**Scene Details:**\n1. **Figurine:** The figure version of the photo I gave you, with a clear representation of PVC material, placed on a round plastic base.\n2. **Packaging:** Behind the figure, there should be a partially transparent plastic and paper box, with the character from the photo printed on it.\n3. **Environment:** The entire scene should be in an indoor setting with good lighting."
}
{
    "command": "coser化",
    "model": "openai-like/gemini-2.5-flash-image-preview",
    "description": "把输入图片变成cosplay照片",
    "preset": "chatgpt",
    "inputPrompt": "Create a realistic cosplay photograph of the character in the image. The cosplayer should be wearing a high-quality costume that accurately replicates the character\\'s outfit. Include appropriate props and background setting that matches the character\\'s universe. Focus on accurate representation of costume details and realistic materials. Draw the picture for me with the background of a comic convention. East-asian face."
}
{
    "command": "mc化",
    "model": "openai-like/gemini-2.5-flash-image-preview",
    "description": "把输入图片变成Minecraft风格",
    "preset": "chatgpt",
    "inputPrompt": "Transform the image into a Minecraft-style character. Create a blocky, pixelated version of the character using Minecraft\\'s visual style. Include appropriate Minecraft environment and elements in the background. The generated entities must be Minecraft-style entities or blocks/structures."
}
1 个赞

没有钱!

3 个赞

八嘎!!!

1 个赞

学宝好强

2 个赞