Skip to main content

Optmyzr Sidekick MCP - User Guide

This article desccribes in detail what the Optmyzr Sidekick MCP is and provides configuration steps.

Written by Anshika Srivastava

Optmyzr provides a Model Context Protocol (MCP) endpoint that allows AI assistants to securely interact with your Optmyzr account through Sidekick.

With MCP, you can connect AI clients such as Claude to Optmyzr and perform the same analysis, reporting, and workflow creation that you would normally do in the Optmyzr UI.

This article explains what MCP is, how it works with Sidekick, and how to configure the Optmyzr MCP endpoint.

What is MCP?

Model Context Protocol (MCP) is a standard that allows AI assistants to connect to external tools and data sources.

Using MCP, an AI client can:

  • Query structured data from external systems

  • Execute workflows and tools

  • Retrieve documentation and contextual information

In the case of Optmyzr, MCP allows supported AI assistants to access Sidekick capabilities.

When Should You Use the Optmyzr MCP?

The Optmyzr MCP is useful when you want to interact with Optmyzr from an external AI assistant instead of working directly inside the Optmyzr interface.

You may find MCP helpful if you want to:

  • Work from your AI workspace. If you use Claude as your primary AI assistant, MCP allows you to analyze performance, generate strategies, and retrieve Optmyzr data without opening the Optmyzr platform.
    ​

  • Combine Optmyzr with other tools. AI assistants can combine Optmyzr data with information from other sources, such as documents, spreadsheets, or analytics platforms, to produce broader analyses.
    ​

  • Build AI-driven workflows. MCP enables AI assistants to chain multiple actions together, such as analyzing performance, generating a Rule Engine strategy, and creating alerts in a single interaction.
    ​

  • Run advanced analysis across accounts. AI assistants can retrieve data from multiple accounts and summarize trends or opportunities across your portfolio.

For many day-to-day optimization tasks, using Sidekick directly inside Optmyzr will still be the simplest option. MCP simply provides another way to access Sidekick capabilities from external AI environments.

Sidekick and MCP

Optmyzr Sidekick is Optmyzr's AI assistant designed to help advertisers analyze performance, discover insights, and build optimization workflows.

Through MCP, the same capabilities available in the Sidekick interface inside Optmyzr can also be accessed through compatible AI clients. This means you can:

  • Run PPC analysis on connected ad accounts

  • Generate Rule Engine strategies

  • Retrieve performance reports

  • Create and review alerts

  • Discover relevant Optmyzr tools

  • Build custom AI workflows that combine multiple tasks

  • Create and apply new positive and negative keywords

All actions executed through MCP interact with your Optmyzr account in the same way as using Sidekick in the UI.

You can get a full overview of Sidekick features here.

Configure the Optmyzr MCP

There are two ways to connect your AI client to Optmyzr MCP:

  • OAuth (recommended): Sign in with your Optmyzr account. No API key needed. Currently supported for Claude and ChatGPT.

  • API key: Generate a key from Optmyzr Settings and pass it when connecting. Supported for all other MCP-compatible clients (Claude Code, Copilot Studio, Langdock, and others).

Connect with OAuth (recommended for Claude and ChatGPT)

The OAuth connection lets you sign in to Optmyzr directly from your AI client — no API key generation or manual configuration required.

Server URL: https://tools.optmyzr.com/OptmyzrMcp

Claude Browser or Desktop

  1. Open Claude Desktop

  2. Go to Settings

  3. Navigate to Connectors

  4. Click Add custom connector

  5. Enter a name (and optional description), then paste the server URL: https://tools.optmyzr.com/OptmyzrMcp

  6. Click Add

  7. Click Connect, sign in to your Optmyzr account, and approve access

Claude will now have access to Optmyzr MCP functions.

On Team and Enterprise Claude plans, only admins or Owners can add connectors to the organisation. Once added, individual users can connect and enable the connector themselves. You can read more about how to add a custom connector to your organisation here.

ChatGPT

  1. Open ChatGPT

  2. Go to Settings, then navigate to Connectors (enable Developer Mode if prompted)

  3. Click Create app (or Create)

  4. Enter a name, paste the server URL: https://tools.optmyzr.com/OptmyzrMcp, and set Authentication to OAuth

  5. Check the box to confirm you trust the app for developer use

  6. Click Create (or Save)

The connector will now be available in your ChatGPT workspace.

Connect with an API Key

For all other MCP clients, connect using an API key generated from your Optmyzr account. Before you configure, generate your MCP API Key from the MCP Integration section of the Settings page in Optmyzr.

Please note that you will have to store the API key securely. It will not be shown once you leave the page. Once generated, you will be able to see when it was created and when it was last used.

You can revoke the API Key, but this will lead any connected MCP clients to lose access.

Claude Browser or Desktop (Manual Configuration)

Prerequisite: Install Node.js

Before using the MCP, you may need to install Node.js on your system.

You can download Node.js directly from the official website: https://nodejs.org/en/download

We recommend using the prebuilt Node.js installer, as it is the fastest and simplest option. You'll only need to select your operating system (Windows, macOS, or Linux), and follow the installation instructions provided.

Once installed, you'll be ready to proceed with MCP setup.

If you are using the free version of Claude, you can configure MCP manually.

  1. Open Claude Desktop

  2. Go to Settings

  3. Navigate to Developer

  4. Click Edit Config

  5. Open the file: claude_desktop_config.json

  6. Add the following configuration:

    {
    "mcpServers": {
    "Optmyzr": {
    "command": "npx",
    "args": [
    "-y",
    "mcp-remote",
    "https://tools.optmyzr.com/OptmyzrMcp",
    "--header",
    "Authorization:YOUR_API_KEY"
    ]
    }
    }
    }

  7. Save the file

  8. Restart Claude Desktop

Claude will now be able to access Optmyzr through MCP.

Claude Code

  1. Open your terminal

  2. Run the following command in your terminal:

    claude mcp add --transport http optmyzr https://tools.optmyzr.com/OptmyzrMcp --header "Authorization:YOUR_API_KEY"

  3. Claude Code will now have access to Optmyzr MCP functions.

Using Claude on a Team or Enterprise plan?

An admin or Owner must add the connector for the whole organisation first. To do this:

  1. Go to Settings → Connectors → Add → Custom → Web

  2. Enter a name and paste the server URL: https://tools.optmyzr.com/OptmyzrMcp

  3. Under Advanced settings, leave OAuth Client ID and Secret empty

  4. Under "How your team connects", select Individual sign-in

  5. Click Add

Once added, individual team members can connect to and enable the connector from their own Claude settings.

You can read more about custom connectors here.

Chat GPT

  1. Open ChatGPT

  2. Go to Settings, then navigate to Apps, and ensure that Developer Mode is enabled.

  3. Click on Create App.

  4. Enter the app name as MCP and provide the following MCP server URL:

    https://tools.optmyzr.com/OptmyzrMcp?api_key=YOUR_API_KEY

  5. Set the authentication type to No Auth

  6. Check the option I understand and want to connect

  7. Click Create, and you will be redirected to the standard developer screen.

  8. To use the Optmyzr MCP, click the + icon in the chat input, go to More, and select Optmyzr MCP from the list.

Other

MCP is an open protocol supported by many clients. Your specific client documentation can advise you on how to connect. Use the server URL https://tools.optmyzr.com/OptmyzrMcp and pass your API key in the Authorization header. For example, a client might accept the following configuration:

{   "optmyzr": {     "url": "https://tools.optmyzr.com/OptmyzrMcp",     "headers": {       "Authorization": "YOUR_API_KEY"     }   } }

Note: Optmyzr's MCP is currently supported on AI platforms that allow MCP integrations.

Google Gemini

At this time, Google Gemini does not natively support MCP integrations. While limited workarounds may exist in developer environments, MCP is not supported in Gemini's standard interface.

Microsoft Copilot

1. Open Copilot Studio and navigate to your Agent
2. Go to Tools → Add Tool → New Tool → MCP
3. Enter the following MCP server URL:

https://tools.optmyzr.com/OptmyzrMcp

4. When prompted for authentication, select API Key and enter your Optmyzr API key
5. Save and your agent will now have access to Optmyzr MCP functions
​
​Note: Generative orchestration must be enabled in your Copilot Studio agent to use MCP. If you are on a Microsoft 365 Copilot plan, MCP support is available through declarative agents — refer to Microsoft's documentation for setup steps.

Langdock

  1. Open Langdock and go to Integrations

  2. Click Add Integration

  3. Enter the following MCP server URL: https://tools.optmyzr.com/OptmyzrMcp

  4. Select API Key as the authentication method

  5. Enter your Optmyzr API key

  6. Click Test connection to verify access and see available tools

  7. Select the tools you want your agents to use

  8. Click Save tools

Your Langdock agents will now have access to Optmyzr MCP functions.

Revoking MCP Access

OAuth connections: Revoke access under "Authorized Applications" in your Optmyzr account settings.

API key connections: Revoke from Settings → MCP Integration.

Note: Revoking the API key will disconnect all clients using it.

Available MCP Functions

The Optmyzr MCP provides the following functions that AI clients can use to interact with your Optmyzr account.

Function

Description

get_active_accounts

Returns all active advertising accounts linked to your Optmyzr user. Accepts an optional search query to filter by account name or ID.

get_ppc_report

Fetches ad account performance data — campaigns, keywords, ads, account-level, or change history. Supports sorting, pagination, and numeric/text filters.

get_competitor_insights

Pulls Google Ads Auction Insights data showing which competitor domains are bidding on the same keywords, with overlap change vs. prior period.

get_competitor_drilldown

Drills into a specific competitor domain (from get_competitor_insights) to see the exact keywords, ad groups, or campaigns you share with them.

get_industry_insights

Benchmarks a Google Ads account's key metrics (CTR, CPC, CVR, Impression Share) against the industry vertical, with percentile rankings and period-over-period changes.

get_account_optimization_recommendations

Analyzes an ad account to surface issues and optimization suggestions grouped by type (keywords, bids, ads, budgets, etc.), sorted by business urgency. Includes a review_url per suggestion.

get_configured_alerts

Fetches alerts set up in Optmyzr for an account. Filterable by alert type (KPI or Budget), level (account, campaign, etc.), and metric.

get_triggered_alerts

Fetches recently fired alerts for an account. Filterable by source type (Anomaly, Rule Engine, Budget Monitor, etc.), level, and metric.

create_or_edit_alert

Creates or edits a KPI/budget alert via a two-step flow: first fetch the schema (get_schema), then submit the alert (submit_alert).

get_merchant_feed_details

Retrieves merchant feed details for a Google Ads or Microsoft Ads account, with optional filtering by feed name.

re_generate_strategy_chain

Generates a complete Rule Engine strategy from a plain-language description. Creates the strategy, executes it, and returns AI-powered recommendations.

re_account_strategy

Builds a rule preview for account-wide or portfolio-wide rules, or change-history review.

re_campaign_strategy

Builds a rule preview for campaigns, ad groups, budgets, or bid strategies.

re_keyword_strategy

Builds a rule preview for keywords and search terms.

re_ad_strategy

Builds a rule preview for individual ads/creatives.

re_audience_strategy

Builds a rule preview for audiences, demographics, or device bid adjustments.

re_geo_strategy

Builds a rule preview for geographic locations (country/region/city/postal code).

re_placement_strategy

Builds a rule preview for display/website placements.

re_shopping_strategy

Builds a rule preview for shopping products, product groups, or Amazon ASINs.

fetch_relevant_tools

Suggests the most relevant Optmyzr tools based on a user query — useful as a starting point when unsure which tool to use.

fetch_help_articles

Searches and returns relevant Optmyzr help articles based on a user query (up to 10 results).

update_language_preference

Updates the language preference for your Optmyzr account using a 2-letter ISO language code.

get_entity_change_schema

Returns required fields for staging a keyword/negative-keyword change; must be called before staging.

preview_entity_changes

Stages new positive/negative keywords into existing ad groups and opens a review card for the user.

edit_previewed_changes

Modifies a still-open preview card (add/update/remove staged rows).

apply_previewed_changes

Applies a reviewed preview card once the user clicks Apply (triggered by the UI, not the assistant).

_internal_get_previewed_changes_state

Internal-only UI state hook; not used directly

Sidekick vs MCP

Both Sidekick and MCP bring AI into your Optmyzr workflow, but they're designed for different kinds of tasks. Here's how you can use each of them.

Use Sidekick when you want to work inside Optmyzr

Sidekick is built directly into the Optmyzr interface and understands exactly what you're looking at. It's your go-to when you want to analyze, learn, and take action within a tool.

Use Sidekick to:

  • Understand what you're seeing on a specific page

  • Get tool-specific recommendations and walkthroughs

  • Apply changes directly (like optimizations, budgets, etc.)

  • Work with features tied closely to the UI (like Keyword Lasso or Rule Engine actions)

  • Continue conversations as you move across tools in your account

In short: Sidekick helps you do the work inside Optmyzr.

Use MCP when you want to work across tools, accounts, or platforms

MCP (Model Context Protocol) connects Optmyzr data to external AI environments (like Claude), letting you analyze and act on data more flexibly, even outside the Optmyzr UI.

Use MCP to:

  • Analyze and compare data across multiple accounts* at once

  • Work with alerts at scale (e.g., review, prioritize, or investigate patterns across accounts)

  • Generate reports or summaries using data pulled from multiple sources

  • Create workflows that go beyond Optmyzr (e.g., draft an email, share updates in Slack, or schedule follow-ups)

  • Create and apply new positive and negative keywords using any AI agent

  • Combine Optmyzr data with other tools in a single workflow

*While Sidekick can help with alerts, summaries, and recommendations within a single account and tool, MCP extends this by letting you go broader and connect workflows across systems.

In short: MCP helps you think, analyze, and orchestrate workflows beyond Optmyzr.

Key Differences at a Glance

Capability

Sidekick

MCP

Works inside Optmyzr UI

✅ Yes

❌ No

Context-aware (based on current page)

✅ Yes

❌ No

Apply changes directly in tools

✅ Yes

✅ Yes

Access UI-specific features (e.g., Keyword Lasso)

✅ Yes

❌ No

Cross-account analysis

⚠️ Limited

✅ Yes

Works with external tools (e.g., Gmail, Slack, etc.).

❌ No

✅ Yes

Automate workflows across platforms

❌ No

✅ Yes

Additional Resources on Optmyzr MCP

Did this answer your question?