Keitaro MCP Server

Keitaro MCP Server

MCP server for Keitaro Tracker Admin API providing 37 tools for campaign management, traffic analytics, reports, and raw click/conversion data, accessible from MCP-compatible AI agents.

Category
访问服务器

README

Keitaro MCP Server

MCP server for Keitaro Tracker Admin API. Provides 37 tools for campaign management, traffic analytics, reports, and raw click/conversion data — accessible from Claude Code, Cursor, and other MCP-compatible AI agents.

Quick Start

Prerequisites

Step 1: Clone the repository

git clone https://github.com/KeitaroManager/keitaro-mcp.git

Choose where to clone. It can be anywhere — your home directory, a tools folder, etc:

# Example locations:
git clone https://github.com/KeitaroManager/keitaro-mcp.git ~/keitaro-mcp
git clone https://github.com/KeitaroManager/keitaro-mcp.git ~/tools/keitaro-mcp
git clone https://github.com/KeitaroManager/keitaro-mcp.git /opt/keitaro-mcp

Step 2: Install the MCP server

Pick ONE of the two scopes below. Replace the placeholder values with your real tracker URL and API key.

Option A — Project scope (available only when working inside a specific project directory):

cd /path/to/your/project

claude mcp add keitaro -s project \
  -e KEITARO_URL=https://your-tracker.example.com \
  -e KEITARO_API_KEY=your-api-key-here \
  -- uv --directory /absolute/path/to/keitaro-mcp run python -m keitaro_mcp

Option B — User scope (available in all your projects, globally):

claude mcp add keitaro -s user \
  -e KEITARO_URL=https://your-tracker.example.com \
  -e KEITARO_API_KEY=your-api-key-here \
  -- uv --directory /absolute/path/to/keitaro-mcp run python -m keitaro_mcp

Step 3: Done

Restart Claude Code. The keitaro_* tools are now available. Try:

You: "List my Keitaro campaigns"
You: "Show ROI by campaign for last 7 days"
You: "Create a campaign named Test"

Full Example (copy-paste ready)

# 1. Clone
git clone https://github.com/KeitaroManager/keitaro-mcp.git ~/keitaro-mcp

# 2. Install (project scope — run this inside your project directory)
claude mcp add keitaro -s project \
  -e KEITARO_URL=https://tracker.mycompany.com \
  -e KEITARO_API_KEY=abc123def456 \
  -- uv --directory ~/keitaro-mcp run python -m keitaro_mcp

# 3. Restart Claude Code and use it

How to Get API Key

  1. Log into your Keitaro tracker admin panel
  2. Go to Maintenance → Users (or click your account icon → Account → API keys)
  3. Click Create API key
  4. Copy the key — it cannot be viewed again after creation

The API key requires admin or expert-level permissions.

Updating

When a new version is released, pull the latest code:

cd /path/to/keitaro-mcp
git pull

That's it. The next time Claude Code starts, it will use the updated version automatically. No reinstall needed — uv reads the source directly.

Uninstalling

# Remove from project scope
claude mcp remove keitaro -s project

# Or remove from user scope
claude mcp remove keitaro -s user

Multiple Keitaro Instances

If you have more than one Keitaro tracker (e.g. production + staging), create a JSON config file:

[
  {
    "name": "production",
    "url": "https://tracker.example.com",
    "api_key": "prod-api-key",
    "description": "Production tracker"
  },
  {
    "name": "staging",
    "url": "https://staging.example.com",
    "api_key": "staging-api-key",
    "description": "Staging tracker"
  }
]

Save it (e.g. ~/keitaro-instances.json), then install with:

claude mcp add keitaro -s project \
  -e KEITARO_CONFIG_FILE=/absolute/path/to/keitaro-instances.json \
  -- uv --directory /absolute/path/to/keitaro-mcp run python -m keitaro_mcp

When multiple instances are configured, tools accept an optional instance parameter. If only one instance exists, it is selected automatically.

Connecting to Cursor

Add to Cursor settings (Settings → MCP Servers → Add):

{
  "keitaro": {
    "command": "uv",
    "args": ["--directory", "/absolute/path/to/keitaro-mcp", "run", "python", "-m", "keitaro_mcp"],
    "env": {
      "KEITARO_URL": "https://your-tracker.example.com",
      "KEITARO_API_KEY": "your-api-key"
    }
  }
}

What It Does

37 Tools by Category

Category Tools Operations
Campaigns 7 List, get, create, update, archive, clone, enable/disable
Streams (Flows) 6 List, get, create, update, delete, enable/disable
Offers 5 List, get, create, update, archive
Landing Pages 5 List, get, create, update, archive
Traffic Sources 3 List, get, create
Affiliate Networks 3 List, get, create
Domains 3 List, get, check DNS/SSL
Groups 1 List by type
Reports 1 Build aggregated report (dimensions, measures, filters, sort)
Clicks 1 Query raw click log with pagination
Conversions 1 Query raw conversion log with pagination
Platform 1 List configured instances

Analytics: Report Builder

The keitaro_build_report tool is the primary analytics interface:

  • 30+ metrics: clicks, conversions, revenue, profit, ROI, CR, CPC, CPA, EPC, cost, leads, sales...
  • 50+ dimensions: campaign, offer, landing, country, day, hour, device, browser, OS, sub_id_1..30...
  • 20+ filter operators: EQUALS, GREATER_THAN, CONTAINS, IN_LIST, BETWEEN, MATCH_REGEXP...
  • Date ranges: any from/to with timezone support

Example:

User: "Show me ROI by campaign for the last 7 days, sorted by profit"
→ keitaro_build_report(date_from="2026-03-17", date_to="2026-03-24",
    dimensions=["campaign"], measures=["clicks", "conversions", "revenue", "profit", "roi"],
    sort=[{name: "profit", order: "DESC"}])

Raw Data: Clicks & Conversions

Query individual click and conversion records with column selection, filtering, and pagination. Useful for debugging traffic quality, checking specific sub_ids, or investigating conversions.

Full Tool List

<details> <summary>Read Tools (safe, no side effects) — 19 tools</summary>

Tool Description
keitaro_list_instances List configured tracker instances
keitaro_list_campaigns List campaigns (supports limit/offset)
keitaro_get_campaign Get campaign by ID
keitaro_list_streams List streams for a campaign
keitaro_get_stream Get stream by ID
keitaro_list_offers List all offers
keitaro_get_offer Get offer by ID
keitaro_list_landing_pages List all landing pages
keitaro_get_landing_page Get landing page by ID
keitaro_list_traffic_sources List all traffic sources
keitaro_get_traffic_source Get traffic source by ID
keitaro_list_affiliate_networks List all affiliate networks
keitaro_get_affiliate_network Get affiliate network by ID
keitaro_list_domains List all domains
keitaro_get_domain Get domain by ID
keitaro_list_groups List groups by type
keitaro_build_report Build aggregated analytics report
keitaro_get_clicks Query raw click records
keitaro_get_conversions Query raw conversion records

</details>

<details> <summary>Write Tools (create, update, modify) — 18 tools</summary>

Tool Description
keitaro_create_campaign Create a new campaign
keitaro_update_campaign Update campaign fields
keitaro_delete_campaign Archive campaign (soft delete, reversible)
keitaro_clone_campaign Duplicate campaign with streams
keitaro_toggle_campaign Enable or disable campaign
keitaro_create_offer Create a new offer
keitaro_update_offer Update offer fields
keitaro_delete_offer Archive offer
keitaro_create_stream Create stream in a campaign
keitaro_update_stream Update stream fields
keitaro_delete_stream Delete stream
keitaro_toggle_stream Enable or disable stream
keitaro_create_landing_page Create landing page
keitaro_update_landing_page Update landing page
keitaro_delete_landing_page Archive landing page
keitaro_create_traffic_source Create traffic source
keitaro_create_affiliate_network Create affiliate network
keitaro_check_domain Check domain DNS/SSL status (triggers active probe)

</details>

Architecture

keitaro-mcp/
├── pyproject.toml                  # Package config, dependencies
├── README.md                       # This file
├── LICENSE                         # MIT
├── .env.example                    # Configuration template
├── keitaro-instances.example.json  # Multi-instance config template
└── src/keitaro_mcp/
    ├── __init__.py                 # Entry point
    ├── __main__.py                 # python -m keitaro_mcp
    ├── server.py                   # MCP Server — 37 tools, match/case routing
    ├── client.py                   # HTTP client (curl subprocess)
    ├── registry.py                 # Multi-instance support
    └── errors.py                   # Error types

Why curl instead of requests/httpx?

Keitaro trackers are typically behind Cloudflare. Python HTTP libraries (urllib, httpx, requests) use TLS fingerprints that Cloudflare identifies and blocks with 403 Forbidden. curl uses a different TLS stack that passes through. This is a deliberate architectural choice.

Design Principles

  • Low-level MCP Server with match/case routing — scales well for 37+ tools
  • Zero Python dependencies beyond mcp>=1.0.0 — uses system curl for HTTP
  • Multi-instance — one server can connect to multiple Keitaro trackers
  • Structured errors — all errors return JSON {"error": "..."}, server never crashes
  • AI-optimized — every tool has detailed description with field lists, enums, and usage hints

Keitaro API Reference

This server wraps the Keitaro Admin API v1. Full OpenAPI spec covers 119 endpoints. This MCP server covers the most commonly used operations.

Key Keitaro concepts:

  • Campaign → has Streams (flows) → each stream routes to Landings and Offers
  • Offers belong to Affiliate Networks (for postback tracking)
  • Traffic Sources define where clicks come from (Facebook, Google, etc.)
  • Domains are tracker URLs used for campaign links
  • Groups organize campaigns, offers, and landings into categories

Contributing

  1. Clone the repo
  2. Install dev dependencies: uv pip install -e ".[dev]" (or pip install -e ".[dev]")
  3. Make changes in src/keitaro_mcp/
  4. Test: python -c "from keitaro_mcp.server import TOOLS; print(f'{len(TOOLS)} tools OK')"
  5. Submit a PR

To add a new tool:

  1. Add the API method to client.py
  2. Add the Tool() definition to TOOLS list in server.py
  3. Add the match case handler in call_tool() in server.py

License

MIT

推荐服务器

Baidu Map

Baidu Map

百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。

官方
精选
JavaScript
Playwright MCP Server

Playwright MCP Server

一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。

官方
精选
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。

官方
精选
本地
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。

官方
精选
本地
TypeScript
VeyraX

VeyraX

一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。

官方
精选
本地
graphlit-mcp-server

graphlit-mcp-server

模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。

官方
精选
TypeScript
Kagi MCP Server

Kagi MCP Server

一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。

官方
精选
Python
e2b-mcp-server

e2b-mcp-server

使用 MCP 通过 e2b 运行代码。

官方
精选
Neon MCP Server

Neon MCP Server

用于与 Neon 管理 API 和数据库交互的 MCP 服务器

官方
精选
Exa MCP Server

Exa MCP Server

模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。

官方
精选