.codex-plugin/plugin.json

{
  "author": {
    "name": "Adzviser LLC"
  },
  "description": "Adzviser is a marketing and sales analytics app for ChatGPT. Connect Google Ads, Meta Ads, GA4, Shopify, Google Search Console, and other supported platforms to retrieve reports and analyze performance in plain language.\n\nCheck advertising spend, campaign conversions, website traffic, search clicks, and store orders. Compare metrics across date ranges and explore results by campaign or other available dimensions, using data from your connected accounts.\n\nSupported data sources:\n\nAdvertising: Google Ads, Meta (Facebook) Ads, Microsoft Ads (Bing), LinkedIn Ads, TikTok Ads, Pinterest Ads, Snapchat Ads, X (Twitter) Ads, Reddit Ads, Amazon Ads, Amazon DSP, Apple Ads, Spotify Ads, AdRoll, Criteo, Instacart, MNTN, Vibe.co, The Trade Desk, Yahoo DSP, Display & Video 360 (DV360), Search Ads 360 (SA360), Campaign Manager 360 (CM360).\n\nAnalytics and search: Google Analytics 4 (GA4), Adobe Analytics, Matomo, Google Search Console, Bing Webmaster Tools, Google Business Profile, Google PageSpeed Insights.\n\nSocial media: Facebook Insights (pages and posts), Instagram Insights (profiles and posts), LinkedIn Company Pages, TikTok Organic, Pinterest Organic, Threads Insights, YouTube Analytics, Sprout Social.\n\nE-commerce: Shopify, WooCommerce, BigCommerce, Amazon Seller Central, Amazon Vendor Central, eBay, TikTok Shop, Google Merchant Center.\n\nCRM, email and call tracking: HubSpot, Salesforce, Zoho CRM, Pipedrive, ActiveCampaign, Klaviyo, Mailchimp, Marketo, Omnisend, CallRail, CallTrackingMetrics.\n\nPublic ad libraries: Google Ads Transparency Center, Meta Ad Library.",
  "interface": {
    "capabilities": [],
    "category": "Data & Analytics",
    "defaultPrompt": [
      "How did our Facebook Ads campaigns perform last month?",
      "How much did we spend on Google Ads last week?",
      "What was our Shopify revenue last week vs the same previous period?"
    ],
    "developerName": "Adzviser LLC",
    "displayName": "Adzviser",
    "longDescription": "Adzviser is a marketing and sales analytics app for ChatGPT. Connect Google Ads, Meta Ads, GA4, Shopify, Google Search Console, and other supported platforms to retrieve reports and analyze performance in plain language.\n\nCheck advertising spend, campaign conversions, website traffic, search clicks, and store orders. Compare metrics across date ranges and explore results by campaign or other available dimensions, using data from your connected accounts.\n\nSupported data sources:\n\nAdvertising: Google Ads, Meta (Facebook) Ads, Microsoft Ads (Bing), LinkedIn Ads, TikTok Ads, Pinterest Ads, Snapchat Ads, X (Twitter) Ads, Reddit Ads, Amazon Ads, Amazon DSP, Apple Ads, Spotify Ads, AdRoll, Criteo, Instacart, MNTN, Vibe.co, The Trade Desk, Yahoo DSP, Display & Video 360 (DV360), Search Ads 360 (SA360), Campaign Manager 360 (CM360).\n\nAnalytics and search: Google Analytics 4 (GA4), Adobe Analytics, Matomo, Google Search Console, Bing Webmaster Tools, Google Business Profile, Google PageSpeed Insights.\n\nSocial media: Facebook Insights (pages and posts), Instagram Insights (profiles and posts), LinkedIn Company Pages, TikTok Organic, Pinterest Organic, Threads Insights, YouTube Analytics, Sprout Social.\n\nE-commerce: Shopify, WooCommerce, BigCommerce, Amazon Seller Central, Amazon Vendor Central, eBay, TikTok Shop, Google Merchant Center.\n\nCRM, email and call tracking: HubSpot, Salesforce, Zoho CRM, Pipedrive, ActiveCampaign, Klaviyo, Mailchimp, Marketo, Omnisend, CallRail, CallTrackingMetrics.\n\nPublic ad libraries: Google Ads Transparency Center, Meta Ad Library.",
    "privacyPolicyURL": "https://docs.adzviser.com/privacy",
    "shortDescription": "Talk to your marketing data",
    "supportURL": "https://adzviser.com/contact-us",
    "termsOfServiceURL": "https://docs.adzviser.com/terms",
    "websiteURL": "https://adzviser.com",
    "logo": "./assets/adzviser-directory-light.png",
    "composerIcon": "./assets/adzviser-directory-light.png"
  },
  "name": "app-6949dd4e2608819189e651a8086079cb",
  "skills": "./skills",
  "version": "5.0.0",
  "mcpServers": "./.mcp.json",
  "extensions": {
    "com.openai": {
      "publication": {
        "countries": [],
        "release_notes": "Updated the app description and guidance for checking ad spend, website traffic, campaign purchases, and store orders."
      },
      "review": {
        "test_cases": {
          "positive": [
            {
              "description": "Google Ads monthly spend. Fresh conversation; Adzviser enabled; reviewer account has access to the Demo workspace and the named source. Use Data Mode. Resolve relative dates when the case is run and state the inclusive dates in the answer. If Demo or its connection is unavailable, report that rather than switching workspaces.",
              "prompt": "How much did we spend on Google Ads last month in the Demo workspace?",
              "tools_triggered": "list_workspace, list_metrics_and_breakdowns_google_ads, retrieve_reporting_data",
              "expected_behavior": "Resolve last month to the complete previous calendar month, not a rolling 30 days. Call retrieve_reporting_data with workspace_name Demo, one date_ranges pair, and assorted_requests.google_ads_request containing metrics [\"Cost\"] and breakdowns [\"Account Name\", \"Account Currency Code\"]. Read the account table, including returned headers such as $ Cost. Report total spend with its returned currency and dates; do not divide an already normalized Cost by one million or combine different currencies. Use only the requested source, omit time_granularity, and answer from structuredContent.reportingData. Surface any note_of_assumption. Do not open a widget or invent zero activity when no rows are returned."
            },
            {
              "description": "GA4 active users for one complete week. Fresh conversation; Adzviser enabled; reviewer account has access to the Demo workspace and the named source. Use Data Mode. Resolve relative dates when the case is run and state the inclusive dates in the answer. If Demo or its connection is unavailable, report that rather than switching workspaces.",
              "prompt": "How many active users did we have last week in GA4 in Demo?",
              "tools_triggered": "list_workspace, list_metrics_and_breakdowns_ga4, retrieve_reporting_data",
              "expected_behavior": "Interpret last week as the previous complete Monday-Sunday week and state that date range. Call retrieve_reporting_data with workspace_name Demo, one inclusive date_ranges pair, and assorted_requests.ga4_request containing metrics [\"Active users\"] and no breakdowns. Report the returned period-wide active-user count. Do not request daily rows and sum them, or substitute Sessions, Total users or the rolling 7-day active users metric. Use only the requested source, omit time_granularity, and answer from structuredContent.reportingData. Surface any note_of_assumption. Do not open a widget or invent zero activity when no rows are returned."
            },
            {
              "description": "Meta purchase conversions by campaign. Fresh conversation; Adzviser enabled; reviewer account has access to the Demo workspace and the named source. Use Data Mode. Resolve relative dates when the case is run and state the inclusive dates in the answer. If Demo or its connection is unavailable, report that rather than switching workspaces.",
              "prompt": "How many purchase conversions did we have per campaign in Meta Ads last month in Demo?",
              "tools_triggered": "list_workspace, list_metrics_and_breakdowns_fb_ads, retrieve_reporting_data",
              "expected_behavior": "Resolve last month to the complete previous calendar month, not a rolling 30 days. Call retrieve_reporting_data with workspace_name Demo and assorted_requests.fb_ads_request containing metrics [\"Total Purchases\"] and breakdowns [\"Campaign ID\", \"Campaign Name\"]. Present one row per returned campaign with its purchase count, keeping distinct campaign IDs even if names match. Label the conversion definition as Total Purchases. Do not substitute Website Custom Conversions, Initiate Checkout, leads, clicks or revenue, or infer the metric from a campaign name. Use only the requested source, omit time_granularity, and answer from structuredContent.reportingData. Surface any note_of_assumption. Do not open a widget or invent zero activity when no rows are returned."
            },
            {
              "description": "Google Search Console monthly clicks. Fresh conversation; Adzviser enabled; reviewer account has access to the Demo workspace and the named source. Use Data Mode. Resolve relative dates when the case is run and state the inclusive dates in the answer. If Demo or its connection is unavailable, report that rather than switching workspaces.",
              "prompt": "How many Google Search Console clicks did we get last month in Demo?",
              "tools_triggered": "list_workspace, list_metrics_and_breakdowns_search_console, retrieve_reporting_data",
              "expected_behavior": "Resolve last month to the complete previous calendar month, not a rolling 30 days. Call retrieve_reporting_data with workspace_name Demo, one inclusive date_ranges pair, and assorted_requests.gsc_request containing metrics [\"Clicks\"] and no breakdowns. Report the returned total clicks for the connected Search Console property. Do not substitute Google Ads clicks or GA4 sessions, and do not add query or page dimensions to a total-only request. Use only the requested source, omit time_granularity, and answer from structuredContent.reportingData. Surface any note_of_assumption. Do not open a widget or invent zero activity when no rows are returned."
            },
            {
              "description": "Shopify monthly order count. Fresh conversation; Adzviser enabled; reviewer account has access to the Demo workspace and the named source. Use Data Mode. Resolve relative dates when the case is run and state the inclusive dates in the answer. If Demo or its connection is unavailable, report that rather than switching workspaces.",
              "prompt": "How many Shopify orders did we have last month in Demo?",
              "tools_triggered": "list_workspace, list_metrics_and_breakdowns_shopify, retrieve_reporting_data",
              "expected_behavior": "Resolve last month to the complete previous calendar month, not a rolling 30 days. Call retrieve_reporting_data with workspace_name Demo, one inclusive date_ranges pair, and assorted_requests.shopify_request containing metrics [\"Orders\"] and no breakdowns. Report the returned order count and period. Do not substitute units sold, customers, revenue or an order-line count. No customer names, email addresses or order-level identifiers are needed. Use only the requested source, omit time_granularity, and answer from structuredContent.reportingData. Surface any note_of_assumption. Do not open a widget or invent zero activity when no rows are returned."
            }
          ],
          "negative": [
            {
              "description": "Unsupported campaign budget change. Explain that Adzviser can report on Google Ads but cannot change campaign budgets. Do not call its reporting, preview or discovery tools as a substitute, and do not claim a change occurred.",
              "prompt": "Increase all my Google Ads campaign budgets in Demo by 20%."
            },
            {
              "description": "Unsupported analytics deletion. Explain that Adzviser cannot delete GA4 data or properties. Do not call this plugin's tools or claim data was deleted.",
              "prompt": "Delete all the data in my GA4 property connected to Demo."
            },
            {
              "description": "Unsupported order refunds. Explain that Adzviser reports on Shopify data but cannot issue refunds or alter orders. Do not call this plugin's tools or claim money was refunded.",
              "prompt": "Refund every Shopify order from last month in Demo."
            }
          ]
        },
        "demo_recording_url": "https://youtu.be/U9YHJKabsF8",
        "commerce": false,
        "commerce_description": "Adzviser subscriptions are purchased separately on the Adzviser website (https://adzviser.com). The plugin retrieves and analyzes connected marketing and sales data; it does not offer in-plugin purchases, checkout, payment processing, ad buying, campaign budget changes, or order refunds."
      }
    }
  },
  "apps": "./.app.json"
}