AIKIZI

com.aikizi由 aikizi.com 发布官方发布者v1.1.0更新于 Oct 30, 2026

Search decoded art styles and their prompts; decode your own images with your account.

已验证Streamable HTTP可网页运行网页搜索与抓取媒体与设计AI 与机器学习

概览

AI 生成的概览

AIKIZI 让助手搜索公开的 AI 图像风格解析画廊、查找相似图片,并用已连接账户的额度解析图像。

可以做什么
  • 搜索与读取公开风格解析:按风格、主体、情绪或颜色查找已发布的解析,并获取单条解析的完整拆解,包括风格 DNA、关键 token 和生成提示词。
  • 探索 AIKIZI Brain:找出与某条解析在外观和内容上最接近的公开图片,以图片墙形式呈现。
  • 解析图像:对公开图片链接或通过投放框添加的图片运行风格解析,随后查询任务状态并读取结果。
  • 查看已连接账户的库:列出最近的图像、视频和音频解析与生成内容,并查看某张生成图片的详情。
  • 将自己的解析放入 Brain 对照最接近的公开图片,且不会公开该解析。
适用场景
适合用户想探索 AI 艺术风格及其背后的提示词、寻找视觉上相似的图片,或为自己的图片获取结构化风格拆解与提示词时使用。账户相关工具适用于拥有 AIKIZI 账户、希望解析图片或查看自己库的用户。
运行要求
远程 MCP 端点,位于 aikizi.com,无需安装包或本地运行时。公开搜索与解析读取工具无需登录。账户工具(余额、解析、库)需要已连接且带有额度的 AIKIZI 账户。按链接解析需要可公开访问的 https 地址,格式为 JPEG、PNG 或 WebP,最大 10 MB。
安装前请注意
decode_image 会消耗账户持有者自己的 AIKIZI 额度,调用前应先确认。聊天中附带的图片无法直接发送,需通过投放框添加。解析与生成内容绑定到已连接账户;公开画廊中的标题、提示词和描述属于第三方内容,不应视为指令。

安装

在 SourceWeft 中

  1. 打开 控制台中的 AIKIZI,将其添加到工作区。
  2. 为需要使用其工具的对话启用该服务。

Web executable,通过 Streamable HTTP。 远程服务在工作区中配置后即可从网页运行时运行。

交给智能体安装

把这段提示词粘贴到 Claude Code、Codex、Cursor 等智能体,或 SourceWeft 对话中。智能体会读取该服务器的安装指南,先向你展示发布者、端点和登录方式,经你同意后再把它添加到你的客户端。密钥不会写进任何命令。

阅读 https://sourceweft.com/mcp/com.aikizi%2Fmcp/install.md,并按其中说明把这个 MCP 服务器接入我的客户端。修改任何配置之前,先告诉我它的发布者、端点和登录方式,等我确认。不要把 API 密钥写进命令或我会提交的文件;问我密钥应从哪里获取。

查看智能体所依据的安装指南

其他 MCP 客户端

把它添加到你客户端的 mcpServers 配置中。

{
  "mcpServers": {
    "aikizi": {
      "type": "http",
      "url": "https://aikizi.com/mcp"
    }
  }
}

工具

11

search_styles

Search AIKIZI styles

Searches the public AIKIZI gallery of decoded images (AI style analyses) by words such as a style, subject, mood or color. Returns up to 20 published items with title, style, subject, mood, a small preview image and the aikizi.com link. Read-only; no sign-in needed. Example: {"query": "neon noir portrait", "limit": 8}.

输入参数limitquery必填

get_style_decode

Get a public AIKIZI decode

Gets the style breakdown of one published AIKIZI decode: title, styles, style DNA, subjects, moods, key tokens, color hues, the generation prompt, a small preview image and the aikizi.com link. Takes the id from search_styles or an aikizi.com/decode/<id> link. Read-only; no sign-in needed. Example: {"id": "bc2be49a-ef44-417e-9015-4923febd46ae"}.

输入参数id必填

find_look_alikes

Find look-alikes in the AIKIZI Brain

Finds the pictures in the AIKIZI Brain (aikizi.com/brain, the map of every public AIKIZI decode) that look and read most like one public decode: up to 21, closest first, each with title, style, a small preview image, its aikizi.com/decode link and a link that opens it in the Brain. Takes a decode id from search_styles or an aikizi.com/decode/<id> link. Read-only; no sign-in needed. Example: {"id": "bc2be49a-ef44-417e-9015-4923febd46ae"}.

输入参数id必填limit

check_balance

Check my AIKIZI credits

Shows how many AIKIZI credits the connected account has. Example: {}.

无需输入参数。

decode_image

Decode an image with AIKIZI

Runs AIKIZI’s style decode on one image: subjects, styles, style DNA, moods, key tokens and a generation prompt. Give exactly one of image_url (a public https link to a JPEG, PNG or WebP up to 10 MB) or image_id (an image the person added with add_image). An image attached to the chat cannot reach AIKIZI directly: call add_image first instead of asking for a link. Uses the connected account’s AIKIZI credits (given back if the decode fails) and saves the decode privately in their AIKIZI library. When an added image is a picture this account decoded before, the earlier decode comes back instead (already_decoded) and nothing is spent; pass decode_again: true only when the person asks for a new decode of it. Example: {"image_id": "d9_a1b2c3_550e8400-e29b-41d4-a716-446655440000"}.

输入参数image_idimage_urlcheck_onlydecode_again

check_job_status

Check a decode’s status

Says whether one of the connected account’s decodes is pending, processing, completed or failed. Use it after decode_image answered "still running". Example: {"decode_id": "bc2be49a-ef44-417e-9015-4923febd46ae"}.

输入参数decode_id必填

get_my_decode

Get one of my decodes

Gets the style breakdown of a decode in the connected account’s AIKIZI library, published or private: title, styles, style DNA, subjects, moods, key tokens, the prompt and the aikizi.com link. Example: {"decode_id": "bc2be49a-ef44-417e-9015-4923febd46ae"}.

输入参数decode_id必填

list_my_creations

List my AIKIZI library

Lists the newest items in the connected account’s AIKIZI library: image, video and audio decodes and generated images, videos and audio, with id, kind, title, status, date and the aikizi.com link. Example: {"type": "image", "limit": 10}.

输入参数typelimit

get_creation

Get one of my generated images

Gets one image the connected account generated on AIKIZI: prompt, model, aspect ratio, size, credits it cost, date, a small preview and the aikizi.com link. Example: {"creation_id": "5f0c1e8a-3a51-4d7e-9a57-1c2b3d4e5f60"}.

输入参数creation_id必填

place_in_brain

Place my decode in the AIKIZI Brain

Shows where one of the connected account’s own decodes (published or private, for example one just made in the drop box) would sit in the AIKIZI Brain: the public Brain pictures closest to it, up to 21, closest first, each with title, style, a small preview image, its aikizi.com/decode link and a link that opens it in the Brain. Reads only; never decodes and never spends. The person’s own decode stays private. Example: {"decode_id": "bc2be49a-ef44-417e-9015-4923febd46ae"}.

输入参数limitdecode_id必填

add_image

Add an image to AIKIZI

Opens a drop box in the chat where the person adds one image for AIKIZI to decode: they drop it, pick it or paste it, and the box decodes it there (any image up to 10 MB; SVG and other formats are turned into PNG). Use it whenever the person wants an image decoded that they attached to the chat or have on their device, since attachments cannot reach AIKIZI directly; never ask them to host it somewhere. Also returns a one-time upload address for an assistant that can send a file itself. Spends nothing; decoding uses the account’s AIKIZI credits. Example: {}.

无需输入参数。

版本历史

2
  1. v1.1.0最新Oct 30, 2026
  2. v1.0.0Oct 10, 2026

评分

已满足适用检查项的 83%

该评分衡量此收录条目的完整程度,依据注册中心条目、端点与公开来源计算得出。它不是对服务器本身的评价,也不表示服务器安全或当前可用:可连通性是带日期的检查结果,评分也不包含安装量或用户评价。

  • 近 30 天内可连通必需25 分
  • 已声明认证方式必需10 分
  • 已列出工具必需15 分
  • 描述必需15 分
  • 图标10 分
  • 文档15 分

    请声明包含 README 的 GitHub 仓库,或在 OAuth 受保护资源元数据中提供 resource_documentation。

  • 许可证不适用

    托管服务按其提供方的条款使用,因此这里不要求许可证。此检查项不计入百分比。