Skip to main content
HireOtto is a set of hosted MCP servers for performance marketers. It gives an MCP-capable AI client structured, permission-bound access to Google Ads, LinkedIn Ads, Google Tag Manager, Google Search Console, and Google Analytics 4. Use it to ask questions, investigate performance, inspect measurement configuration, and—where the connected server supports it—prepare or apply reviewable changes from the same conversation. You do not need to build a developer app, configure a cloud project, manage API credentials, or run a local server.
HireOtto is an execution layer, not an autonomous media buyer. Your AI client can use the tools and data you authorize, but the marketer still owns the objective, business context, interpretation, and approval of consequential changes.

​
What HireOtto connects

Each HireOtto server exposes a defined set of tools. Access is limited by the HireOtto plan, the marketing-platform identity you connect, and that identity’s permissions. Google Ads and Search Console share one HireOtto endpoint, but Google treats them as separate services. Connecting Google Ads does not grant Search Console access; authorize Search Console separately when you need organic-search data.
A server’s write capability is not permission to make every possible platform change. HireOtto supports specific actions, while Google Ads and LinkedIn still enforce account roles, object state, policy, validation, and other platform rules. Review the exact account, object, current value, and proposed value before approving a consequential edit.

​
How it works

MCP, or Model Context Protocol, is a standard way for AI clients to use external tools and data. HireOtto runs the MCP server remotely and handles the platform connection. The workflow has two authorization layers:
  1. Connect the AI client to HireOtto. Add the relevant remote MCP endpoint and complete the HireOtto sign-in.
  2. Connect the marketing platform. From the AI client, start the Google or LinkedIn authorization flow and approve access with an identity that can open the resources you need.
  3. Verify the resource. Run a small read request and confirm the returned account, property, container, role, currency, and other identifiers before deeper analysis or any write.
  4. Ask for the job. Describe the business question or task in normal marketing language. The AI client chooses an available HireOtto tool, supplies its parameters, and returns the result in the conversation.
  5. Review the result. Check the evidence and scope. For supported write workflows, inspect the proposed change and verify the saved state afterward.
Your Google or LinkedIn password is never pasted into the conversation. OAuth grants HireOtto only the access approved during authorization, and the underlying platform continues to enforce that identity’s permissions.

​
Server endpoints

Add only the connections you need: Your AI client must support remote MCP servers over HTTP and OAuth. HireOtto works with MCP-capable clients and workflows including Claude, ChatGPT, Cursor, VS Code, and Make. Menu names and workspace-level availability can differ by client. For client-specific steps, see Connect HireOtto to an AI client.

​
Start with a read request

A small verification prompt confirms both the MCP connection and the platform authorization. It also helps prevent work in the wrong account or property.

List the Google Ads accounts I can access. Include each account name, customer ID, currency, time zone, and whether it is a manager account. Do not change anything.

List the LinkedIn Ads accounts I can access. Include each account ID, currency, my role, and whether that role allows supported writes. Do not change anything.

List the Google Tag Manager accounts and web containers I can access. Include account IDs, container IDs, public IDs, and container names. Do not change anything.

List the Google Search Console properties I can access. Show the exact property URL, property type, and my permission level. Do not change anything.

List the Google Analytics 4 accounts and properties I can access. Include account IDs, property IDs, display names, time zones, and currencies. Do not change anything.

If the expected resource is missing, check the connected platform identity before changing the prompt. HireOtto cannot discover an account, container, or property that the authorized identity cannot access.

​
From questions to reviewable work

HireOtto is most useful when the prompt defines the decision you need, the scope to inspect, and the boundary on what may change.

​
Analyze one platform

For Google Ads customer 1234567890, compare the previous 28 complete days with the prior 28 days. Explain the material changes in spend, conversions, CPA, and conversion value by campaign. Separate evidence from interpretation, identify missing context, and do not change anything.

​
Inspect measurement configuration

Review the live published Google Tag Manager container for our website. Inventory conversion-related tags, their firing and blocking triggers, paused state, and important parameters. Flag ambiguous or missing wiring, but do not claim that a tag fired in the browser and do not change anything.

​
Combine evidence across servers

Cross-platform analysis requires every relevant HireOtto server to be connected in the same AI workflow. Results from different platforms should stay clearly labelled because their metrics, attribution, scopes, and data freshness are not interchangeable.

Investigate the decline in paid conversions over the previous 28 complete days. Use Google Ads for campaign and conversion performance, GA4 for acquisition and on-site outcomes, and Google Tag Manager for the current published conversion-tag configuration. Keep each platform's evidence separate, state date and attribution differences, label confirmed versus possible causes, and do not change anything.

​
Prepare a controlled change

For supported write workflows, separate investigation from execution. Ask the AI client to show the current value, proposed value, affected object, and expected scope before anything changes.

In Google Ads customer 1234567890, inspect campaign 987654321 and show its current daily budget, status, bidding strategy, and recent performance. Propose one budget change with the current and proposed values, explain the evidence, and wait for my explicit approval before applying anything.

After an approved write, ask the AI client to read the live object again. A successful action response is not a substitute for verifying the saved state.

​
Plans, credits, and profiles

All current plans include supported core access to Google Ads, LinkedIn Ads, Tag Manager, Search Console, and GA4 beta. Simple discovery requests generally use fewer credits than deeper reports, planning requests, or multi-step work. Agency-only audit tools combine several checks into one repeatable workflow. Starter, Pro, and Agency do not set plan-based limits on the number of supported platforms, ad accounts, or managed spend; usage on Starter and Pro still consumes credits. Each connection uses a default platform profile unless a profile is specified. Multiple connected profiles are an Agency capability. A profile represents a separate authorized login; assigning another label to the same login does not expand its platform access. See Credits, billing, and plans and the Feature and entitlement matrix for current details.

​
Defaults and limits

HireOtto provides opinionated defaults so common marketing requests can start quickly, but there is no single date range, row limit, or output default shared by every server and tool. The selected tool determines its supported parameters, default reporting window, row limit, output modes, and credit cost. Before a high-stakes or large request:
  • State the exact account, property, container, campaign, or other resource identifier.
  • Use complete dates and name the comparison period.
  • Specify the business metric that should guide interpretation.
  • Request CSV output when the relevant tool supports exports and the complete row set matters.
  • Check the server’s tools reference for tool-specific defaults, maximum ranges, row limits, and export expiry.
  • Split a large investigation into smaller requests if the AI client, platform API, or tool returns partial or truncated data.
Platform limits still apply. Reporting can be delayed, estimated, sampled, thresholded, privacy-filtered, or unavailable for low-volume segments. Search Console and advertising platforms may omit data; GA4 dimensions and metrics are not always compatible; a GTM configuration inventory cannot prove browser runtime behavior; and an audience-size estimate is not a performance forecast.

​
Common failure cases

​
What HireOtto does not decide

HireOtto can make platform data and supported actions available to an AI client. It does not know your sales feedback, lead quality, margin, legal constraints, creative standards, risk tolerance, or business priorities unless you provide them. Treat the AI client’s output as a structured starting point:
  • Evidence should remain distinguishable from interpretation.
  • Recommendations should name the account and entity they affect.
  • Low-volume results should be labelled as uncertain.
  • Cross-platform comparisons should preserve each platform’s definitions.
  • Consequential changes should have an explicit approval boundary.
  • Every applied write should be read back from the platform.

​
Choose your next step

Connect a server

Add the right endpoint, authorize the platform, and verify your first resource.

Google Ads quickstart

Connect Google Ads and run a safe first account request.

LinkedIn Ads quickstart

Verify Campaign Manager access, role, and write capability.

Tag Manager quickstart

Inspect containers and tag wiring without changing GTM.

Search Console quickstart

Authorize organic-search access and verify properties.

GA4 quickstart

Connect the read-only GA4 beta and verify properties.