Examples

Copy-paste MCP examples for CloudX reporting and documentation search, including real argument payloads and prompt patterns.

Examples

These examples use the real tool names exposed by CloudX MCP. Reporting examples use exact UTC timestamps.

Example 1: Dashboard Summary For US iOS Traffic

Question:

Show me dashboard performance for US iOS production traffic from April 15, 2026 through April 21, 2026. Summarize revenue, fill rate, CTR, and daily trend.

Tool call:

{
  "name": "GetReportDashboard",
  "arguments": {
    "start_time": 1776211200,
    "end_time": 1776815999,
    "country": "US",
    "device_os": "iOS",
    "test_mode": "production"
  }
}

What you can expect back:

a summary object for totals and derived KPIs
a chart_data array with daily rows for the selected period

Example 2: Compare Bidders On Android Test Traffic

Question:

For April 15, 2026, compare bidder performance on Android test traffic in the US. Rank bidders by revenue and call out bid rate and win rate.

Tool call:

{
  "name": "GetReportBidders",
  "arguments": {
    "start_time": 1776211200,
    "end_time": 1776297599,
    "country": "US",
    "device_os": "Android",
    "test_mode": "test"
  }
}

Typical fields:

name
requests
bids
bid_rate
impressions
win_rate
revenue
ecpm

Question:

Show banner ad unit performance for com.example.game in the US on Android for April 15, 2026. I want to find the weakest placements by fill rate.

Tool call:

{
  "name": "GetReportAdUnits",
  "arguments": {
    "start_time": 1776211200,
    "end_time": 1776297599,
    "country": "US",
    "device_os": "Android",
    "test_mode": "all",
    "app_bundle": "com.example.game",
    "ad_unit_type": "BANNER"
  }
}

Use this pattern when you already know the app bundle or want one ad format only.

Example 4: Export Rows For Analysis

Question:

Export CloudX report rows for US Android traffic from April 1, 2026 through April 7, 2026 so I can inspect the raw table structure.

Tool call:

{
  "name": "GetReportExport",
  "arguments": {
    "start_time": 1775001600,
    "end_time": 1775606399,
    "country": "US",
    "device_os": "Android"
  }
}

This tool returns a JSON shape with:

columns
rows
row_count

Example 5: Search CloudX Documentation

Question:

Find CloudX docs about report filters and summarize which filters are available.

Tool call:

{
  "name": "SearchDocs",
  "arguments": {
    "body": {
      "query": "report filters"
    }
  }
}

This tool returns matching docs pages with:

content
path
url
metadata

Example 6: Build A Custom Country Breakdown

Question:

Break down US and international performance by country for April 15, 2026. Return requests, impressions, and revenue, ranked by the top 20 countries by revenue.

Tool call:

{
  "name": "GetReportBreakdown",
  "arguments": {
    "start_time": 1776211200,
    "end_time": 1776297599,
    "by": "country",
    "metrics": "requests,impressions,revenue",
    "test_mode": "production",
    "top": 20
  }
}

The response identifies the requested dimensions and metrics, then returns grouped values in rows.

Example 7: Inspect An Auction And Its Bids

Start by finding recent auctions:

{
  "name": "GetAuctionList",
  "arguments": {
    "start_time": 1776211200,
    "end_time": 1776297599,
    "app_bundle": "com.example.game",
    "has_ilrd": true,
    "limit": 10
  }
}

Then pass an auction_id from the result to GetAuctionShow:

{
  "name": "GetAuctionShow",
  "arguments": {
    "auction_id": "auction_123",
    "with": "rounds,bids,ilrd,external-ilrd"
  }
}

Use GetAuctionRounds for aggregate round metrics and GetAuctionBids when you need to search bid rows across many auctions.

Example 8: Review A/B Test Performance

{
  "name": "GetReportABTest",
  "arguments": {
    "ab_test_id": "abtest_123",
    "start_time": 1776211200,
    "end_time": 1776815999
  }
}

Compare control and test cumulative and daily metrics. Treat statistical fields as unavailable when they are omitted or null.

Example 9: Edit, Validate, And Publish Configuration

First inspect the live configuration with GetConfigShow and use the returned resource IDs. This example changes one ad unit’s floor in a draft:

{
  "name": "PostConfigEdit",
  "arguments": {
    "body": {
      "summary": "Raise the banner floor",
      "operations": [
        {
          "update_ad_unit": [
            {
              "id": "adunit_123",
              "bidfloor": 1.5
            }
          ]
        }
      ]
    }
  }
}

Ordinary edits return a draft id and validation result. Review that draft directly:

{
  "name": "GetConfigShow",
  "arguments": {
    "id": "draft_123"
  }
}

Validate it again if needed, then publish only after review:

{
  "name": "PostConfigPublish",
  "arguments": {
    "body": {
      "draft_id": "draft_123",
      "version_label": "Raise the banner floor"
    }
  }
}

Prompting Tips

ask one reporting question per tool call when possible
use exact dates such as April 15, 2026 instead of last week
specify country, device_os, and test_mode when you need a narrow answer
use GetReportExport when you want rows, not a narrative summary
use GetReportDashboard first when you need a fast overall health check
use GetReportBreakdown when the fixed report tools do not provide the grouping you need
use GetAuctionList before GetAuctionShow when you do not know the auction ID
ask the client to show and validate a config draft before calling PostConfigPublish
use SearchDocs first when your question is about CloudX setup, configuration, or feature behavior