Claude Desktop MCP Stopped Working: How to Fix It

On May 1st, your Claude Desktop MCP connector stopped working. You did not change anything. This guide explains why your MCP broke silently. It shows how to diagnose the root cause with the Google Analytics MCP as an example. It also shows how to prevent future breakage with version pinning.

What you need before starting

• Claude Desktop (current version installed on your Mac or Linux machine)
• An active Google Cloud project with OAuth 2.0 desktop app credentials (if fixing the GA MCP specifically)
• Access to your Claude config file at ~/Library/Application Support/Claude/claude_desktop_config.json
• Basic familiarity with YAML syntax and environment variables
• 5–10 minutes for diagnosis and fix (longer if creating new OAuth credentials)

Understanding why MCP connectors break silently

If you install an MCP server with uvx and the @latest tag, uvx does not download a pinned version. Instead, uvx checks PyPI every time Claude Desktop starts and fetches the latest published version.

Third-party MCP packages (for example mcp-google-analytics) live on PyPI, not on your machine. Each time you start Claude, uvx queries PyPI for a newer version and installs it without notice. If the new version changes the API, adds new scopes, or changes authentication, your old credentials and configuration can stop working.

This is not a Claude Desktop update. Claude itself does not change. A third-party developer published a new version of their MCP package. uvx @latest installed it without notice to you.

On May 1st, a new version of mcp-google-analytics added features that need the Google Analytics Admin API in addition to the Data API. The Admin API needs the analytics.edit scope. The existing refresh token had only the analytics.readonly scope. When Claude called the Admin API method, Google’s OAuth server rejected the request: ACCESS_TOKEN_SCOPE_INSUFFICIENT (403).

Diagnosing a broken MCP: three checks

Check 1: Look for error messages in Claude Desktop’s console

Open Claude Desktop. If your MCP is broken, the chat or the debug output can show an error. The most common signature is:

ACCESS_TOKEN_SCOPE_INSUFFICIENT

This error means the access token your MCP uses does not have permission for the task.

Check 2: Verify the refresh token has the right scopes

For the Google Analytics MCP, your refresh token must include the analytics.readonly scope. If you created the token before the MCP update, and the new version added Admin API calls, the token is now insufficient.

You cannot see token scopes directly in your config file. Google’s OAuth server stores them. To verify the scopes, create a new token with an explicit scope declaration.

Check 3: Check the MCP package version in your config

Open ~/Library/Application Support/Claude/claude_desktop_config.json and find the MCP server entry. Look for the args line:

"args": ["mcp-google-analytics@latest"]

If you see @latest, your MCP can update automatically without warning. This setting is the root cause of silent breakage.

Fixing the GA MCP connector: step-by-step

Step 1: Create new OAuth credentials in Google Cloud

Go to Google Cloud Console. Create a new project, or use an existing one.

Go to APIs & Services → Credentials. Click + Create Credentials → OAuth client ID → Desktop app.

Download the JSON file. It contains your client_id and client_secret.

Step 2: Re-authenticate with the correct scopes

Run this command in your terminal:

python3 -c "
import webbrowser
import json
from google_auth_oauthlib.flow import InstalledAppFlow

# Update these paths to your credential file
CREDS_FILE = '/path/to/your/client_secret_*.json'
SCOPES = ['https://www.googleapis.com/auth/analytics.readonly']

flow = InstalledAppFlow.from_client_secrets_file(CREDS_FILE, SCOPES)
creds = flow.run_local_server(port=8080)

# Extract refresh token
print('Refresh token:')
print(creds.refresh_token)
"

A browser window opens. Sign in with the Google account that owns your Analytics property. Approve access when the prompt appears. The script prints your new refresh token. Copy this value.

Why analytics.readonly and not analytics.edit? The GA Data API (for data queries) needs only read access. You do not need the Admin API unless you manage GA properties, accounts, or create new properties. For most users, analytics.readonly is enough.

Step 3: Update your Claude config with new credentials

Open ~/Library/Application Support/Claude/claude_desktop_config.json in your text editor. Find your MCP server entry. If the entry is missing, create one:

"google-analytics-thesciencetalk": {
  "command": "/Users/yourusername/.local/bin/uvx",
  "args": ["mcp-google-analytics@0.0.3"],
  "env": {
    "GOOGLE_ANALYTICS_CLIENT_ID": "YOUR_CLIENT_ID.apps.googleusercontent.com",
    "GOOGLE_ANALYTICS_CLIENT_SECRET": "YOUR_CLIENT_SECRET",
    "GOOGLE_ANALYTICS_REFRESH_TOKEN": "YOUR_REFRESH_TOKEN",
    "GOOGLE_ANALYTICS_PROPERTY_ID": "YOUR_PROPERTY_ID"
  }
}

Replace:

• YOUR_CLIENT_ID — from your downloaded JSON (the client_id field)
• YOUR_CLIENT_SECRET — from your downloaded JSON (the client_secret field)
• YOUR_REFRESH_TOKEN — the token you printed in Step 2
• YOUR_PROPERTY_ID — your Google Analytics GA4 property ID (for example 465530700)

Note the version pinning. This example uses @0.0.3 instead of mcp-google-analytics@latest. This choice is intentional. See the next section.

Step 4: Restart Claude Desktop and test

Close Claude Desktop completely. Reopen it. In your chat, try a simple query:

Get me the active users and sessions for the last 7 days.

If the MCP is fixed, you see data. If ACCESS_TOKEN_SCOPE_INSUFFICIENT appears again, check that you created the refresh token with the correct scopes. Also check that you restarted Claude after you updated the config.

Preventing future breakage: pin your package version

The root cause of the May 1st breakage was @latest. Every time Claude starts, uvx queries PyPI. If a new version exists, uvx installs it without notice. If that new version has breaking changes, your MCP breaks without warning.

Replace @latest with a specific version number.

Instead of:

"args": ["mcp-google-analytics@latest"]

Use:

"args": ["mcp-google-analytics@0.0.3"]

What version should I use? Check PyPI for the latest stable version. As of May 2026, 0.0.3 is stable. Replace 0.0.3 with the current version.

Trade-offs:

For production use or daily automation, pin the version. The extra step is worth the reliability, even if it delays access to new features.

Troubleshooting

Error: ACCESS_TOKEN_SCOPE_INSUFFICIENT

Cause: Your refresh token does not have the analytics.readonly scope. Fix: Create a new token. Follow Step 2 above. Make sure the SCOPES list includes https://www.googleapis.com/auth/analytics.readonly.

Error: Invalid Credentials or Unauthorized

Cause: Your CLIENT_ID, CLIENT_SECRET, or PROPERTY_ID is incorrect. Fix: Check each value in your config against your Google Cloud Console. Restart Claude after you make changes.

Error: Property not found

Cause: Your PROPERTY_ID does not exist, or it is not linked to your Google account. Fix: Log into Google Analytics and verify the property ID. Find it in Admin → Property Settings → Property ID.

Error: MCP still not working after restart

Cause: Claude caches the config and can fail to reload it right away. Fix: Close Claude completely. Wait 5 seconds. Reopen Claude. If possible, avoid a force-quit. A clean shutdown helps the config reload.

What you can do now

If you pinned the version, your GA MCP is now stable and does not break without warning. You can now:

• Query Google Analytics data directly in Claude: “What’s my top traffic source this week?”
• Build automation that reliably fetches GA data as part of a larger workflow
• Set up regular reports with scheduled Claude sessions that pull GA metrics

Next steps: If you have multiple GA properties, add them as separate MCP server entries in your config. Use a different env section for each property. If you use other MCP tools from PyPI (for example Google Sheets or WordPress), apply the same version pinning pattern to prevent similar silent breakage.

Related reading on The Science Talk

This guide accompanies the complete GA4 MCP setup guide on The Science Talk, which walks through the initial installation step by step from scratch. That post assumes no prior MCP experience. This post focuses on troubleshooting and on preventing breakage after initial setup.

Want more guides like this? Browse all AI Guides or get in touch →