Google Home Developer MCP Server

The Google Home Developer Model Context Protocol (MCP) server gives AI-powered development tools the ability to search documentation and retrieve real-time, domain-specific information:

• Google Home Developer Center public documentation

• OpenThread
• Matter Specification: Available for version 1.5.1
• Thread Specification: Available for version 1.4.1

In addition, the Home Developer MCP server provides user feedback integration diagnostics for Cloud-to-cloud smart home integrations to improve user experience based on user feedback, whereas the Google Home Vitals dashboard is designed for generic quality monitoring. User feedback integration diagnostics correlate end-user feedback reports and categorize integration issues with diagnostic logs:

• Anomaly distribution: Summarize symptom distribution across user feedback issues using query_user_feedback_anomaly_distribution.
• Historical trends: Query daily issue counts grouped by symptom over specific time ranges using query_user_feedback_symptom_trend.
• Issue discovery: Retrieve sample issue summaries and affected devices using fetch_user_feedback_sample_issues.
• Diagnostic analysis: Inspect detailed execution logs and Home Graph state histories using batch_fetch_user_feedback_sample_issue_logs.

An MCP server acts as a proxy between an external service that provides context or data and an AI application (the host). The Home Developer MCP server exposes executable tools that let Large Language Models (LLMs) find precise answers directly from indexed corpuses. To learn more about MCP servers, see What is the MCP and how does it work?.

Prerequisites

To use the Home Developer MCP server, you must have an MCP-compatible AI client or code editor installed. See Choose an authentication path to select the proper authentication method for the content and features you want to access.

Google Antigravity

If you are using Google's AI development environment, install one of the following:

• Antigravity CLI, following the Getting Started with Antigravity CLI guide
• Antigravity 2.0, following the Google Antigravity Documentation
• Antigravity IDE, following the Antigravity IDE Getting Started guide

Third-party AI assistants

You can also integrate the MCP with other AI applications, command line tools, and code editors, such as:

• Claude Code
• Cursor
• GitHub Copilot in VS Code

User feedback integration diagnostics requirements

Through the Home Developer MCP server, you can access user feedback integration diagnostics. To retrieve this data, you must meet the following requirements:

• You must have access to a Google Cloud project with the Home Developers API (homedevelopers.googleapis.com) enabled.
• You must authenticate your MCP client using OAuth 2.0 or Application Default Credentials (ADC) signed in with your authorized Google Account for viewing the diagnostics for your integration projects.

Choose an authentication path

The Home Developer MCP server supports three authentication modes depending on whether you need quick trial access, dedicated production quota for public documentation, or access to partner insights and gated Partner Program documentation.

Unauthenticated connection testing

This method does not require an API key, OAuth client, or Google Cloud project. It uses a shared public quota and returns publicly available documentation only, making it the fastest way to trial the server and test basic tool calls.

API key authentication for public documentation

To prevent rate-limiting under shared public quotas when querying public documentation, you can authenticate with a Google Cloud API key.

1. In the Google Cloud console, enable the API on your project:
   1. Open the Home Developers API page in the Google APIs library.
   2. Check that you have the correct project selected in which you intend to use the API.
   3. Click Enable.

2. Create an API Key:

   1. Go to the Credentials page.
   2. Click Create credentials, and then select API key from the menu.
   3. In the Name field, provide a name for the key.

   4. Click the Select API restrictions drop-down, and then type Home Developers API. Click the result, and then click OK.

   5. Click Create.

   6. After the key is created, click Show key to view it, and copy it for the next step.

OAuth registration and setup

When you authenticate with OAuth 2.0 or Application Default Credentials (ADC), the Home Developer MCP server resolves your Google Account identity from the OAuth bearer token and verifies your partner integration permissions and Google Group memberships in real time. This authentication mode unlocks:

• User feedback integration diagnostics, enabling partners to analyze symptom anomalies, inspect historical trends, and retrieve diagnostic logs for their Cloud-to-cloud integrations.

All OAuth requests to the Home Developer MCP server require two configurations:

• The Home Developers API (homedevelopers.googleapis.com) must be enabled on your Google Cloud project.
• The OAuth credential must be authorized by your Google Account and include the https://www.googleapis.com/auth/cloud-platform scope (which MCP OAuth clients and gcloud auth application-default login request by default).

Complete the following steps to enable the API and set up either an OAuth 2.0 client or Application Default Credentials (ADC):

1. Enable the Home Developers API on your Google Cloud project:

   1. Open the Home Developers API page in the Google Cloud console, select your Google Cloud project, and click Enable.

   2. Alternatively, enable the service using the Google Cloud CLI (gcloud):

      gcloud services enable homedevelopers.googleapis.com \
        --project=YOUR_PROJECT_ID
      

2. Choose and configure your OAuth credential method:

   OAuth 2.0 client

   Use an OAuth 2.0 client ID and client secret when connecting AI clients that manage the OAuth 2.0 flow directly (such as Google Antigravity, Claude, or Cursor).

   1. Configure the OAuth consent screen in Google Auth Platform:
   2. Open the Google Auth Platform Overview page (or APIs & Services > OAuth consent screen) in the Google Cloud console.
   3. If your project doesn't have an OAuth brand configured, click Get started:
      1. Under App Information, enter an App name (for example, Home Developer MCP) and select your User support email, then click Next.
      2. Under Audience, select External (or select Internal if all users belong to your Google Workspace organization), then click Next.
      3. Under Contact Information, enter your contact email address, click Next, accept the terms, and click Create.
   4. Authorize your Google Account on the consent screen:
      1. Open the Audience page under Google Auth Platform.
      2. If Publishing status is set to Testing, click Add users under Test users, enter the exact Google Account email address you use for your Google Home integration or Developer Center sign-in, and click Save. Alternatively, click Publish app under Publishing status.
   5. Register an OAuth 2.0 client ID:
   6. Open the Clients page under Google Auth Platform (or APIs & Services > Credentials).
   7. Click Create client (or + Create credentials > OAuth client ID).
   8. Select the Application type that matches your AI environment:
      • Select Desktop app for local desktop IDEs and CLI agents (such as Cursor, Claude Code, or VS Code) that listen on a local loopback redirect URI (http://localhost or http://127.0.0.1).
      • Select Web application for cloud or web-callback clients, and under Authorized redirect URIs, click Add URI to register your client's callback URL:
        • Google Antigravity: https://antigravity.google/oauth-callback
        • Claude: https://claude.ai/api/mcp/auth_callback and https://claude.com/api/mcp/auth_callback
   9. Enter a Name (for example, Home Developer MCP Client) and click Create.
   10. Copy the Client ID and Client secret shown in the dialog (or click Download JSON) to use in the next section. You don't need to configure OAuth scopes in the Google Cloud console.

   ADC

   Use Application Default Credentials (ADC) if you are connecting from Google Antigravity using google_credentials, or if your MCP client passes an Authorization: Bearer header generated by the gcloud CLI.

   1. Install the Google Cloud CLI (gcloud).
   2. Run the following command to authenticate with your authorized Google Account and configure your quota project:

   gcloud auth application-default login \
     --project=YOUR_PROJECT_ID
   

   1. When your browser opens, sign in with your authorized Google Account and click Allow to grant the requested permissions.
   2. If your client requires a raw bearer token in the Authorization HTTP header rather than reading local ADC credentials directly, print an access token with the following command and pass your project ID in the X-Goog-User-Project header:

   gcloud auth application-default print-access-token
   

Configure your client

Enter your credentials in the following fields to populate the configuration templates for your preferred AI development environment:

You can locate these values in the Google Cloud console:

• Find your Google Cloud Project ID in the project selector drop-down at the top of the Google Cloud console (in the ID column), or on the Cloud Hub overview page under Project ID.
• Find your OAuth Client ID and OAuth Client Secret on the Clients page under Google Auth Platform.
• Find your API key on the Credentials page under APIs & Services.

Enter your Google Cloud Project ID: YOUR_PROJECT_ID
Enter your OAuth Client ID: YOUR_CLIENT_ID
Enter your OAuth Client Secret: YOUR_CLIENT_SECRET
Enter your API key: YOUR_API_KEY

{
  "mcpServers": {
    "HomeDevelopers": {
      "httpUrl": "https://homedevelopers.googleapis.com/mcp",
      "authProviderType": "google_credentials",
      "oauth": {
        "scopes": [
          "https://www.googleapis.com/auth/cloud-platform"
        ]
      },
      "headers": {
        "X-Goog-User-Project": "YOUR_PROJECT_ID"
      }
    }
  }
}

{
  "mcpServers": {
    "HomeDevelopers": {
      "serverUrl": "https://homedevelopers.googleapis.com/mcp",
      "oauth": {
        "clientId": "YOUR_CLIENT_ID",
        "clientSecret": "YOUR_CLIENT_SECRET"
      }
    }
  }
}

{
  "mcpServers": {
    "HomeDevelopers": {
      "serverUrl": "https://homedevelopers.googleapis.com/mcp",
      "headers": {
        "X-Goog-Api-Key": "YOUR_API_KEY"
      }
    }
  }
}

{
  "mcpServers": {
    "HomeDevelopers": {
      "serverUrl": "https://homedevelopers.googleapis.com/mcp"
    }
  }
}

claude mcp add HomeDevelopers \
  --transport http https://homedevelopers.googleapis.com/mcp \
  --header "Authorization: Bearer $(gcloud auth application-default print-access-token)" \
  --header "X-Goog-User-Project: YOUR_PROJECT_ID"

claude mcp add HomeDevelopers \
  --transport http https://homedevelopers.googleapis.com/mcp \
  --header "X-Goog-Api-Key: YOUR_API_KEY"

claude mcp add HomeDevelopers \
  --transport http https://homedevelopers.googleapis.com/mcp

{
  "mcpServers": {
    "HomeDevelopers": {
      "url": "https://homedevelopers.googleapis.com/mcp",
      "auth": {
        "clientId": "YOUR_CLIENT_ID",
        "clientSecret": "YOUR_CLIENT_SECRET"
      }
    }
  }
}

{
  "mcpServers": {
    "HomeDevelopers": {
      "url": "https://homedevelopers.googleapis.com/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_ACCESS_TOKEN",
        "X-Goog-User-Project": "YOUR_PROJECT_ID"
      }
    }
  }
}

{
  "mcpServers": {
    "HomeDevelopers": {
      "url": "https://homedevelopers.googleapis.com/mcp",
      "headers": {
        "X-Goog-Api-Key": "YOUR_API_KEY"
      }
    }
  }
}

{
  "mcpServers": {
    "HomeDevelopers": {
      "url": "https://homedevelopers.googleapis.com/mcp"
    }
  }
}

{
  "servers": {
    "HomeDevelopers": {
      "url": "https://homedevelopers.googleapis.com/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_ACCESS_TOKEN",
        "X-Goog-User-Project": "YOUR_PROJECT_ID"
      }
    }
  }
}

{
  "servers": {
    "HomeDevelopers": {
      "url": "https://homedevelopers.googleapis.com/mcp",
      "headers": {
        "X-Goog-Api-Key": "YOUR_API_KEY"
      }
    }
  }
}

{
  "servers": {
    "HomeDevelopers": {
      "url": "https://homedevelopers.googleapis.com/mcp"
    }
  }
}

Verify installation and execution

Once your tool configuration is updated, verify the connection by prompting your agent workspace.

Verification example in Antigravity CLI

1. Launch your interactive workspace session:

   agy
   

   

2. Execute the discovery checklist command to inspect active MCP servers:

   /mcp
   

3. Confirm that HomeDevelopers (or your chosen server name key) appears in the active connections list.

   

4. Ask a Google Home development question, for example:

   How do I create an Automation using Home APIs?
   

5. When Antigravity requests permission to call the Home Developer MCP server's search_documents tool, select any of the Allow options to continue.

   

6. After calling the tool, Antigravity answers your query with a grounded response from the documentation.

   

Video tutorial

Watch the video for a step-by-step guide to installing the Home Developer MCP.

User feedback integration diagnostics

For Cloud-to-cloud smart home integrations, the Home Developer MCP server provides user feedback integration diagnostics. It correlates end-user feedback reports with Cloud-to-cloud logs.

The server provides four specialized diagnostic tools:

• query_user_feedback_anomaly_distribution: Summarizes the distribution of symptom types across user feedback sample issues for an integration.
• query_user_feedback_symptom_trend: Retrieves daily trend counts grouped by symptom over a specified time window.
• fetch_user_feedback_sample_issues: Retrieves sample issue summaries and affected device records without verbose logs.
• batch_fetch_user_feedback_sample_issue_logs: Fetches detailed Cloud-to-cloud execution logs and Home Graph state histories for specific sample issue IDs.

These tools identify your integration using an agentId, which is your Google Cloud project ID (the same definition in action.devices.SYNC).

Diagnostic investigation workflow example

AI assistants connected to the Home Developer MCP server can orchestrate an end-to-end troubleshooting workflow:

1. Identify top integration pain points: Ask your AI assistant to check the symptom anomaly distribution for your integration over a specific date range. For example:

   Check the user feedback anomaly distribution for agent YOUR_AGENT_ID
   between 2026-09-01 and 2026-09-20.
   

   The assistant calls query_user_feedback_anomaly_distribution to reveal which symptom type accounts for the highest share of user complaints.

2. Evaluate historical trends: Inspect whether a symptom represents an acute regression or an ongoing pattern by querying daily counts:

   Show the daily trend for agent YOUR_AGENT_ID
   over the last 14 days.
   

   The assistant calls query_user_feedback_symptom_trend to plot or summarize daily counts and detect spikes following recent updates.

3. Inspect affected sample issues: Fetch sample issue summaries to identify affected devices:

   Fetch recent sample issues for agent YOUR_AGENT_ID
   over the last 14 days.
   

   The assistant calls fetch_user_feedback_sample_issues to retrieve sample records.

4. Batch fetch diagnostic logs: Retrieve complete logs for the identified sample:

   Retrieve the diagnostic logs for sample IDs SAMPLE_ID.
   

   The assistant calls batch_fetch_user_feedback_sample_issue_logs to further inspect the logs for samples listed in the previous step.

Test with direct HTTP curl requests

For pipeline testing outside of an AI client environment, you can send standard JSON-RPC 2.0 payloads directly to https://homedevelopers.googleapis.com/mcp using curl.

Discover active system tools

Use the tools/list method to inspect the tools exposed by the server:

curl --location "https://homedevelopers.googleapis.com/mcp" \
  --header 'Content-Type: application/json' \
  --header 'Accept: application/json, text/event-stream' \
  --data '{
      "jsonrpc": "2.0",
      "method": "tools/list",
      "id": 1
  }'

Execute a functional search query

Select a tab to test search_documents using OAuth, an API key, or an unauthenticated request:

VALID_TOKEN=$(gcloud auth application-default print-access-token)
curl --location "https://homedevelopers.googleapis.com/mcp" \
  --header 'Content-Type: application/json' \
  --header 'Accept: application/json, text/event-stream' \
  --header "Authorization: Bearer ${VALID_TOKEN}" \
  --header "X-Goog-User-Project: YOUR_PROJECT_ID" \
  --data '{
      "jsonrpc": "2.0",
      "method": "tools/call",
      "id": 2,
      "params": {
          "name": "search_documents",
          "arguments": {
              "query": "How to create an automation?"
          }
      }
  }'

curl --location "https://homedevelopers.googleapis.com/mcp" \
  --header 'Content-Type: application/json' \
  --header 'Accept: application/json, text/event-stream' \
  --header "X-Goog-Api-Key: YOUR_API_KEY" \
  --data '{
      "jsonrpc": "2.0",
      "method": "tools/call",
      "id": 2,
      "params": {
          "name": "search_documents",
          "arguments": {
              "query": "How to create an automation?"
          }
      }
  }'

curl --location "https://homedevelopers.googleapis.com/mcp" \
  --header 'Content-Type: application/json' \
  --header 'Accept: application/json, text/event-stream' \
  --data '{
      "jsonrpc": "2.0",
      "method": "tools/call",
      "id": 2,
      "params": {
          "name": "search_documents",
          "arguments": {
              "query": "How to create an automation?"
          }
      }
  }'

Query user feedback integration diagnostic tools

To test user feedback integration diagnostic tools directly with curl with an OAuth request:

VALID_TOKEN=$(gcloud auth application-default print-access-token)
curl --location "https://homedevelopers.googleapis.com/mcp" \
--header 'Content-Type: application/json' \
--header 'Accept: application/json, text/event-stream' \
--header "Authorization: Bearer ${VALID_TOKEN}" \
--header "X-Goog-User-Project: YOUR_PROJECT_ID" \
--data '{
    "jsonrpc": "2.0",
    "method": "tools/call",
    "id": 3,
    "params": {
        "name": "query_user_feedback_anomaly_distribution",
        "arguments": {
            "agentId": "YOUR_AGENT_ID",
            "firstDate": "2026-09-01",
            "lastDate": "2026-09-20"
        }
    }
}'

VALID_TOKEN=$(gcloud auth application-default print-access-token)
curl --location "https://homedevelopers.googleapis.com/mcp" \
--header 'Content-Type: application/json' \
--header 'Accept: application/json, text/event-stream' \
--header "Authorization: Bearer ${VALID_TOKEN}" \
--header "X-Goog-User-Project: YOUR_PROJECT_ID" \
--data '{
    "jsonrpc": "2.0",
    "method": "tools/call",
    "id": 4,
    "params": {
        "name": "query_user_feedback_symptom_trend",
        "arguments": {
            "agentId": "YOUR_AGENT_ID",
            "firstDate": "2026-09-01",
            "lastDate": "2026-09-20"
        }
    }
}'

VALID_TOKEN=$(gcloud auth application-default print-access-token)
curl --location "https://homedevelopers.googleapis.com/mcp" \
--header 'Content-Type: application/json' \
--header 'Accept: application/json, text/event-stream' \
--header "Authorization: Bearer ${VALID_TOKEN}" \
--header "X-Goog-User-Project: YOUR_PROJECT_ID" \
--data '{
    "jsonrpc": "2.0",
    "method": "tools/call",
    "id": 5,
    "params": {
        "name": "fetch_user_feedback_sample_issues",
        "arguments": {
            "agentId": "YOUR_AGENT_ID",
            "firstDate": "2026-09-01",
            "lastDate": "2026-09-20"
        }
    }
}'

VALID_TOKEN=$(gcloud auth application-default print-access-token)
curl --location "https://homedevelopers.googleapis.com/mcp" \
--header 'Content-Type: application/json' \
--header 'Accept: application/json, text/event-stream' \
--header "Authorization: Bearer ${VALID_TOKEN}" \
--header "X-Goog-User-Project: YOUR_PROJECT_ID" \
--data '{
    "jsonrpc": "2.0",
    "method": "tools/call",
    "id": 6,
    "params": {
        "name": "batch_fetch_user_feedback_sample_issue_logs",
        "arguments": {
            "sampleIds": ["SAMPLE_ID"]
        }
    }
}'

Send feedback

Use our feedback form to submit your thoughts and suggestions about the Home Developer MCP or how we can improve your experience with the tool.

Send us your feedback