Leadpipe

io.github.leadpipe-com由 GitHub 帳號 leadpipe-com 發布v0.2.1更新於 Oct 25, 2026

Official MCP server for Leadpipe intent and data APIs.

概覽

AI 產生的概覽

Leadpipe 官方 MCP 伺服器,讓助理透過 Leadpipe 的意圖與資料 API 探索主題、建立與管理受眾,並查看訪客資料與追蹤像素。

可以做什麼
探索與比較主題,查看面向、趨勢與變動,並分析網站主題。預覽與查詢受眾,列出、建立、更新或刪除已儲存的受眾。查看受眾狀態、結果、執行紀錄、統計與匯出。查詢訪客資料,列出、建立或更新追蹤像素。查看帳戶健康與額度狀態。
適用情境
適合行銷或成長情境:需要主題研究、受眾規模預估與已儲存受眾管理,或訪客與像素診斷,並希望由助理取代 Leadpipe 主控台完成這些操作。
執行需求
以 stdio 在本機執行,通常透過 npx 使用已發布的 npm 套件(需 Node.js)。需要 LEADPIPE_API_KEY 環境變數中的 Leadpipe API 金鑰;LEADPIPE_BASE_URL 為選用,用於非預設環境。需要連線至 Leadpipe API 的網路存取。
安裝前請注意
API 金鑰是放在環境變數中的機密,避免寫入共用設定。工具可建立、更新與刪除已儲存的受眾,以及建立或更新像素,會變更帳戶資料並可能消耗額度;批次操作前先查看帳戶狀態。

安裝

在 SourceWeft 中

  1. 開啟 儀表板中的 Leadpipe,將其新增到工作區。
  2. 為需要使用其工具的對話啟用該服務。

Desktop only,透過 STDIO。 STDIO 服務會啟動本機處理程序,因此需要 SourceWeft 桌面主機。

其他 MCP 客戶端

參照 儲存庫 中的啟動說明。

README

Leadpipe MCP

Official MCP server for Leadpipe intent and data APIs.

Use this server to give agents direct access to:

  • topic discovery
  • website topic analysis
  • audience preview and ad hoc query
  • saved audience management
  • audience status, results, runs, stats, and exports
  • visitor data lookup
  • tracking pixel management
  • account health and credit status

What This Is

Leadpipe MCP wraps the public Leadpipe intent and data APIs in a local MCP server that agents can use over stdio.

This server is API-key authenticated by default so the full audience workflow, visitor data access, and pixel management workflows are available to agents.

Use With Claude, Cursor, and Codex

Most users do not need to clone this repo or build from source. The published package is:

bash
npx -y @leadpipe/mcp

You will need:

bash
export LEADPIPE_API_KEY=sk_...

Optional advanced override:

bash
export LEADPIPE_BASE_URL=https://api.aws53.cloud

LEADPIPE_BASE_URL only needs to change if you are pointing at a non-default environment.

Claude Desktop

Add this to your Claude Desktop MCP config:

json
{  "mcpServers": {    "leadpipe": {      "command": "npx",      "args": ["-y", "@leadpipe/mcp"],      "env": {        "LEADPIPE_API_KEY": "sk_..."      }    }  }}

Cursor

Add this to ~/.cursor/mcp.json:

json
{  "mcpServers": {    "leadpipe": {      "command": "npx",      "args": ["-y", "@leadpipe/mcp"],      "env": {        "LEADPIPE_API_KEY": "sk_..."      }    }  }}

Codex

Add the server with the Codex CLI:

bash
codex mcp add leadpipe --env LEADPIPE_API_KEY=sk_... -- npx -y @leadpipe/mcp

Then confirm it is configured:

bash
codex mcp list

Quick Start

If you just want to verify the package runs:

bash
npx -y @leadpipe/mcp

Run From Source

For local development or repo hacking:

bash
npm installnpm run buildnode dist/index.js

For iterative local development:

bash
npm run dev

Local source build MCP config:

json
{  "mcpServers": {    "leadpipe": {      "command": "node",      "args": ["/absolute/path/to/mcp/dist/index.js"],      "env": {        "LEADPIPE_API_KEY": "sk_..."      }    }  }}

Tool Surface

Topic Discovery

  • list_topics
  • get_topic_facets
  • search_topics
  • get_topic_trend
  • compare_topics
  • get_topic_movers
  • analyze_website_topics

Audience Builder

  • get_audience_filters
  • preview_audience
  • query_audience
  • list_audiences
  • get_audience
  • create_audience
  • update_audience
  • delete_audience

Audience Results

  • get_audience_status
  • get_audience_results
  • list_audience_runs
  • get_audience_stats
  • export_audience

Visitor Data

  • query_visitor_data

Pixels

  • list_pixels
  • create_pixel
  • update_pixel

Account

  • get_account_status

Prompts

  • discover-audience-topics
  • operate-saved-audience
  • investigate-visitor-data

Prompt Examples

See examples/prompt-examples.md for copy-paste prompts that work well with this MCP server.

Good starting prompts:

  • "Find 5-10 B2B topics for cloud cost optimization software, compare their recent trends, and preview an audience."
  • "Create a saved audience for these topics, activate it, wait until it is ready, and then summarize the first page of results."
  • "Check the status of audience <audience-id>. If it is ready, export it and give me the download URL."
  • "Investigate visitor activity for example.com over the last 30 days and tell me whether our tracking setup looks healthy."
  • "List our pixels, find any paused ones, and explain what needs attention."

Agent Playbooks

For more structured operational flows, see:

Resources

  • leadpipe://docs/auth
  • leadpipe://docs/workflows
  • leadpipe://docs/data-api

Typical Agent Workflow

  1. Search or analyze topics
  2. Preview audience size
  3. Create or update a saved audience
  4. Check status until the audience is ready
  5. Fetch results or export CSV

Typical Data Workflow

  1. Check get_account_status
  2. Query visitor activity with query_visitor_data
  3. Inspect or create pixels with list_pixels / create_pixel
  4. Pause or reactivate tracking with update_pixel

Development

bash
npm run typechecknpm run build

License

MIT

來源:README.md,提交 b537df5

工具

0
SourceWeft 尚未在沙箱中啟動過該軟體包,因此其工具未知。

版本歷史

1
  1. v0.2.1最新Sep 16, 2026

評分

已滿足適用檢查項的 50%

該評分衡量此收錄條目的完整程度,依據註冊中心條目、端點與公開來源計算得出。它不是對伺服器本身的評價,也不表示伺服器安全或目前可用:可連通性是帶日期的檢查結果,評分也不包含安裝量或使用者評價。

  • 近 30 天內可連通25 分

    請保持端點能夠回應 MCP 握手;SourceWeft 每 30 天探測一次。

  • 已宣告認證方式10 分

    請在註冊中心條目中宣告憑證請求標頭,或保持端點能夠回應探測,以便判定其認證方式。

  • 已列出工具15 分

    請讓伺服器在無憑證時回應 tools/list,或發布可在沙箱中無憑證啟動的軟體包。

  • 描述必需15 分
  • 圖示10 分
  • 文件15 分
  • 授權條款10 分