QuerySail: the Google Analytics 4 MCP server for ChatGPT, Claude and Codex
A hosted, read-only MCP server for your GA4 data. Ask about users, sessions, channels, landing pages, key events and revenue across every property you can see, in the AI app you already use.
Independent product, not made by Google. You sign in to QuerySail through our sign-in provider, WorkOS, and grant Analytics access on Google’s own consent screen.
What you can ask
Plain questions. Your assistant picks the dates, dimensions, metrics and filters, and every answer says what it covers.
Which Google Analytics properties can you access?
How many users and sessions did my site have in the last 28 days?
Which channels brought the most sessions last month, and how did that change from the month before?
What are my top landing pages by key events?
How many people are on the site right now, and from which countries?
Compare organic search sessions with the same weeks last year.
Capabilities
What it does
Property discovery. Lists the GA4 properties your connected Google accounts can read, grouped by Analytics account.
Any report. Every Data API dimension and metric, including your custom definitions and key events, with filters and sorting.
Comparisons and realtime. What changed between two periods and who is on the site right now.
Clear coverage. Responses state the dates and time zone used, whether more rows exist, and when Google thresholded or sampled the report.
What it does not do
It never writes to Google: no editing properties, key events, audiences or settings.
It cannot read properties your Google account cannot read, or Universal Analytics properties.
It does not store an analytics archive. Data is fetched when you ask, with a short-lived cache.
It does not explain why numbers changed. Your assistant interprets them.
Five read-only MCP tools
All are marked read-only for MCP clients.
list_properties
The GA4 properties your connected Google accounts can read, with their exact IDs such as properties/123456789.
list_fields
The dimensions and metrics a property supports, including its custom definitions, so your assistant uses the right names.
run_report
Any metrics by any dimensions for a date range, with filters, sorting, totals and 20 to 100 rows per call with paging.
compare_report
Two periods side by side: the previous period, the same weekdays a year earlier or your own dates, with the top gaining and losing channels, pages or countries.
run_realtime_report
Active users, views and events in the last 30 minutes, by country, device, page or event.
Set up in about five minutes
No Google Cloud project and nothing to install. Included in the same plan as Search Console; see pricing and usage limits.
Connect your AI app to the MCP server
You need a QuerySail account with Google Analytics connected. Start the free trial, connect Google Analytics under Account → Connections, then add this endpoint to your app and sign in with your QuerySail account when it asks:
https://mcp.querysail.com/ga4/mcp
Open Settings → Security and login and turn on Developer mode. Availability depends on your plan and workspace settings.
claude mcp add --transport http --scope user querysail-ga4 https://mcp.querysail.com/ga4/mcp
Then run /mcp in Claude Code, choose querysail-ga4 and sign in with your QuerySail account.
Leave out --scope user to add it to the current project only.
In Claude Desktop or on claude.ai, open Customize → Connectors, select + Add, then Add custom connector.
Name it QuerySail Analytics, paste the endpoint above as the server URL and select Continue.
Keep the detected sign-in settings, add the connector and sign in with your QuerySail account.
In a chat, select + → Connectors and turn on QuerySail.
On Team and Enterprise plans an owner adds the connector under Organization settings → Connectors, then each member selects Connect. Free plans allow one custom connector.
Add this to ~/.cursor/mcp.json for all projects, or .cursor/mcp.json in one project. If the file already lists servers, add the entry inside the existing mcpServers.
MCP (Model Context Protocol) is the standard way AI assistants call external tools. A Google Analytics MCP server exposes your GA4 reports as tools, so ChatGPT, Claude or Codex can fetch users, sessions, engagement, key events and revenue when you ask a question, instead of you building reports in the Analytics interface.
Do I need a Google Cloud project, service account or API key?
No. You approve read-only Google Analytics access with Google once, on the QuerySail website. Your AI app then signs in with your QuerySail account. There is nothing to install and no key to paste.
Can it change my Analytics settings?
No. QuerySail only asks Google for read-only Analytics access. It cannot create or edit properties, data streams, key events, audiences or any other setting.
What data can my assistant get?
Any Google Analytics 4 dimension and metric the Data API offers, including your custom dimensions, custom metrics and key events: users, sessions, views, engagement, events, conversions and revenue, broken down by date, channel, source and medium, landing page, page, country, device, event and more, with filters. Comparisons show what changed between two periods and which channels or pages gained or lost. Realtime reports cover the last 30 minutes. Universal Analytics properties are not supported.
Does QuerySail store my Analytics data?
Reports are fetched from Google when your assistant asks and kept in a short-lived cache of 15 minutes to avoid repeat calls. Realtime reports are not cached. There is no long-term archive.
How much does it cost?
Google Analytics is included in the same $8 per month plan as Search Console, with unlimited connected properties. The 7-day free trial needs no card and starts with your first successful report in either product.
Is QuerySail made by Google?
No. QuerySail is an independent product that uses the public Google Analytics Data and Admin APIs. It is not affiliated with or endorsed by Google, OpenAI or Anthropic.