How to Connect Google Maps to Claude (and What It Can't Do)
The Google Maps story changed in late 2025, when Google announced managed MCP support for its services and Maps got a dedicated offering: Maps Grounding Lite, a Google-hosted MCP endpoint at mapstools.googleapis.com that grounds any MCP client — Claude included — in data on 300M+ places. So the honest mid-2026 answer to “does Claude work with Google Maps” is: yes, through Google’s official server or a community one, provided you bring a Google Cloud project. There’s still no Google Maps app in Claude’s connector directory, and the connection still lives only inside an open chat.
The five tools in Google’s Maps MCP endpoint
Grounding Lite is deliberately small. Per Google’s MCP reference, it exposes:
| Tool | What it does |
|---|---|
search_places | Find businesses, addresses, and landmarks, with optional location bias |
compute_routes | Routes between two points — driving and walking only, no transit or cycling |
lookup_weather | Current conditions plus hourly (120h) and daily (10-day) forecasts |
resolve_names | Turn names/addresses into canonical place IDs, 20 per batch |
resolve_maps_urls | Extract place IDs from shared Google Maps links, 20 per batch |
If you need classic Maps Platform surface that Grounding Lite skips — geocoding proper, Distance Matrix, elevation — community servers such as cablate/mcp-google-map wrap those APIs directly with your Maps Platform key.
Hooking it up to Claude
Enable the Maps Grounding Lite API in a Google Cloud project and create credentials (API key or OAuth). Then attach the endpoint: in Claude Code, add it as a remote MCP server; from claude.ai, it goes in as a custom connector under Settings → Connectors — which requires a paid Claude plan. Mind Google’s attribution terms too: grounded Maps results are meant to be displayed with attribution, which matters if Claude’s answers feed anything customer-facing.
A connected chat handles requests like:
- “Find coffee shops within a ten-minute walk of 660 Market St and give me their place IDs.”
- “Compute the driving route from SFO to our Palo Alto office — how long, and what’s the weather there this afternoon?”
- “Here are 15 Google Maps links from the team offsite doc — resolve them and list the actual venues.”
A lookup engine, not a dispatcher
Notice what every one of those prompts has in common: you’re present, asking. That’s the whole operating model. Claude with Maps is superb at interactive lookups and useless as infrastructure, because connectors have no way to run when a chat isn’t open. The workflows people actually want Maps data inside — a new order arrives, so geocode the delivery address and send the route to the driver; a lead comes in, so attach travel time from the nearest office to the CRM record — never involve anyone sitting in a chat. They’re event-driven, and no Claude connector fires on an event. The place lookup that takes Claude two seconds on request simply never happens on its own.
Making Maps a step in a workflow: Carly
That dispatcher role is what Carly plays. Carly is an AI executive assistant whose workflows start from triggers — an incoming order, a form submission, a schedule — and run in the cloud end to end. Google Maps becomes one step among several: the trigger fires, Carly resolves the address, pulls the route or travel time, then carries the result forward into a sent email (Gmail or Outlook), a Slack post, a spreadsheet row, or a task. Building it is conversational — say “when an order comes in, look up the address and email the route to dispatch,” answer Carly’s follow-up questions, done. No Cloud project, no key rotation, nothing to host. AI agents start at $35/month, non-AI steps run free and unlimited, and Maps sits alongside 200+ other integrations.
Claude’s Maps MCP vs Carly for location work
| Claude (Maps Grounding Lite MCP) | Carly | |
|---|---|---|
| Place search and place-ID resolution | Yes | Yes |
| Routes with weather context | Yes (drive/walk only) | Yes |
| Geocode each new order as it arrives | No | Yes |
| Email a route to dispatch automatically | No | Yes |
| Works without an open chat window | No | Yes (cloud) |
| Prerequisites | Google Cloud project + credentials, paid Claude plan | None beyond signup |
| Pricing | Maps usage billing + Claude plan | AI agents from $35/mo |
Frequently Asked Questions
Does Claude integrate with Google Maps?
Yes — most cleanly through Google’s official Maps Grounding Lite MCP server at mapstools.googleapis.com, which any MCP client can use with a Google Cloud project and credentials. It’s not a directory app, so on claude.ai it’s added as a custom connector on a paid plan. Community servers cover additional Maps Platform APIs.
What can Claude do with Maps Grounding Lite?
Search places, resolve names and shared Maps URLs to place IDs, compute driving and walking routes, and pull weather forecasts. Transit and cycling routes aren’t included, and batch tools cap at 20 items per call.
Do I need my own Google Maps API key?
Yes, in effect — Grounding Lite requires a Google Cloud project with the API enabled and either an API key or OAuth credentials, billed under your Maps Platform account. Claude doesn’t provide Maps access out of its own pocket.
Can Claude geocode addresses from incoming orders automatically?
No — that’s an event-driven job, and Claude connectors only operate inside conversations a person starts. Carly runs it as a triggered workflow: order arrives, address resolved, route delivered to whoever needs it. AI agents start at $35/month.
Why not just paste an address into regular Claude?
Plain Claude will guess from training data — fine for “roughly where is this,” risky for travel times or current business details. The MCP connection grounds answers in live Maps data; a triggered workflow through Carly grounds them and removes you from the loop.
More: Claude connectors · Claude vs Carly · Can Claude send emails · Claude + Google Sheets · Claude + Google Drive · Claude + Amplitude · Claude + Mixpanel · Claude + Segment
Ready to automate your busywork?
Carly schedules, researches, and briefs you—so you can focus on what matters.
See what people say
"Before Carly, I relied on a Calendly link, but the whole process felt impersonal and not very professional. Carly changed that by handling all the back-and-forth, so I'm no longer stuck in endless email threads trying to line up schedules.
Now Carly reaches out to candidates, shares my real-time availability, lets them pick a slot, then sends a Zoom link and drops it straight into my calendar. She sends reminders to both of us before each call, which has significantly reduced no-shows and last-minute confusion.
On top of scheduling, Carly acts like a full executive assistant, sending me my schedule the night before so I can prepare for each call. It reminds me of the old x.ai assistant, but Carly is noticeably smarter, faster, and better suited to my healthcare recruitment business."
