Lotus — AI Citation Intelligence

app.clicon由 clicon.app 发布官方发布者v1.27.0更新于 Nov 2, 2026

Measure how AI assistants cite your brand. Returns measured data and ready-to-apply fixes.

已验证Streamable HTTP可网页运行数据与分析商业与电商AI 与机器学习

概览

AI 生成的概览

Lotus 衡量 AI 助手是否引用你的品牌，并给出带可安装代码的优先修复项，并在你的线上站点上验证。

可以做什么
  • 对任意域名做 GEO 诊断，给出可见度评分与预估收入风险
  • 列出待处理的快速修复项，并按 hash 获取可直接安装的代码
  • 查看竞争对手动向、生成的产物、AI 爬虫活动与预估业务损失
  • 查看引用评分及逐条查询的测量数据，并核验产物是否已在线上生效
  • 在线上验证通过后将修复项标记为已应用，生成分享链接并重新生成产物
适用场景
适合想了解大模型回答中是否引用自己品牌而非竞争对手，并希望编码助手直接应用 JSON-LD 或 llms.txt 修复、再到生产环境确认的场景。
运行要求
托管远程端点，使用 streamable-http，无需安装。需要 clicon.app 注册流程中签发的 Lotus API 密钥，通过 X-API-Key 请求头发送。仅 analyze_geo 预览无需密钥。
安装前请注意
API 密钥属于凭据，只能放在 X-API-Key 请求头中，不要放进工具参数或网址。写入类工具会改变账户状态：mark_applied、generate_share_link 和 regenerate_artifacts，后者有频率限制且仅限付费套餐。服务器会抓取你的线上站点做验证，但不会写入你的站点或代码仓库。

安装

在 SourceWeft 中

  1. 打开 控制台中的 Lotus — AI Citation Intelligence，将其添加到工作区。
  2. 连接 X-API-Key 凭据。它们会被私密存储，绝不会在公开页面上展示。
  3. 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/app.clicon%2Flotus/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": {
    "lotus-ai-citation-intelligence": {
      "type": "http",
      "url": "https://lotus.clicon.app/mcp/",
      "headers": {
        "X-API-Key": "<YOUR_API_KEY>"
      }
    }
  }
}

README

Lotus — AI Citation Intelligence (MCP Server)

Lotus is a Generative Engine Optimization (GEO) platform. It measures whether your brand is actually cited as a source inside LLM-generated answers, and hands your coding agent the exact code to fix what is missing.

This repository documents the hosted Lotus MCP server. The server is a managed remote endpoint — there is nothing to install or self-host.

  • Endpoint: https://lotus.clicon.app/mcp/ (trailing slash required)
  • Transport: streamable-http, stateless
  • Registry: app.clicon/lotus
  • Website: https://clicon.app

What it does

Search engines told you when you ranked badly. Generative engines do not. If an assistant hallucinates your product, ignores you, or cites a competitor instead, you lose the business with no signal that it happened.

Lotus closes that loop in three steps:

  1. Measure — real citation measurements across LLM answers for your domain, plus crawler activity and Search Console signals.
  2. Fix — prioritized "quick wins" with ready-to-install code (JSON-LD structured data, llms.txt).
  3. Verify — Lotus fetches your live site and confirms the fix is really there. A fix is not "applied" because an agent said so; it is applied because it was found in production.

Tools

analyze_geo works without credentials, so any agent can evaluate a domain before signing up. Every other tool requires an API key. Tools that take no domain argument resolve the domain from the key itself.

Preview

ToolDescription
analyze_geoGEO diagnosis for any domain: visibility score and estimated revenue at risk

Read (API key)

ToolDescription
get_quick_winsPending fixes, each with title, category, priority, estimated impact and a hash
get_quick_win_codeFull, install-ready code for one quick win, by hash
get_competitor_actionsCompetitor moves classified by AI-discoverability layer
get_artifactRetrieve a generated artifact (json_ld, llms_txt)
get_bot_activityAI crawler activity on your site over the last N days
get_bleed_modelEstimated business lost through the generative channel, with its provenance
verify_installationCheck whether your artifacts are live on your site right now
get_citation_verdictLatest citation score plus the per-query measurements behind it

Write (API key)

ToolDescription
mark_appliedMarks a quick win as applied — after verifying it against your live site
generate_share_linkCreate a shareable report link
regenerate_artifactsRegenerate your artifacts. Rate-limited; paid tiers only

Configuration

Add the server to any MCP-compatible client (Claude Code, Claude Desktop, Cursor, VS Code, or your own agent):

json
{  "mcpServers": {    "lotus": {      "url": "https://lotus.clicon.app/mcp/",      "headers": {        "X-API-Key": "YOUR_LOTUS_API_KEY"      }    }  }}

The API key is sent as the X-API-Key header — never as a tool argument or a query parameter. Keys are issued during onboarding at clicon.app. Without a key the server still responds to analyze_geo.

Exact config syntax varies slightly per client; the shape above is what most of them expect.

Usage

Talk to your agent in plain language — it picks the tools:

"Run the Lotus GEO analysis on my domain."

"Show me the pending Lotus quick wins."

"Get me the code for the highest-priority one and apply it to this repo."

"I deployed. Mark it as applied in Lotus."

On that last step Lotus fetches your live site and looks for the schema. If it is there, the fix turns green with a verification date. If it is not — because the deploy has not landed, or the snippet went in wrong — Lotus refuses to mark it. No false positives.

Lotus generates and verifies. You deploy. The server never writes to your site or your repository.

Support

Questions or access: https://clicon.app

来源：README.md，提交 e573bd3

工具

0
SourceWeft 尚未读取此服务器的工具。

版本历史

1
  1. v1.27.0最新Sep 16, 2026

评分

已满足适用检查项的 85%

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

    请让服务器在无凭据时响应 tools/list，或发布可在沙箱中无凭据启动的软件包。

  • 描述必需15 分
  • 图标10 分
  • 文档15 分
  • 许可证10 分