Gemini Data Connectors in 2026: Built-In, Third-Party and MCP Options
Gemini Enterprise offers up to 138 prebuilt connectors, mostly for documents, tickets and chat: Google Drive, Jira Cloud, Slack, GitHub, ServiceNow. Google Ads, Meta Ads and GA4 have no native connector, so marketing data needs a third-party tool.
Third-party Gemini data connectors take different routes. Some sit in Google's catalog as partner connectors. Others plug in as a custom MCP server that your admin registers once. The route affects which edition you need, how fresh the numbers are, and who maintains the setup.
This guide explains how Gemini data connectors work, what the native catalog includes by edition, and how seven native and third-party options compare for marketing and sales reporting. It ends with setup steps for both prebuilt and custom connectors.
How connectors in Gemini work
Google Gemini connectors come in two separate systems, one for each Gemini product. The Gemini app is for personal use. Gemini Enterprise is a paid workplace platform that an admin manages, formerly known as Google Agentspace until the rebrand in October 2025. Gemini AI connectors in one product don't carry over to the other.
Connected Apps in the Gemini app
In the Gemini app, connectors are called Connected Apps, and you manage them under Customize > Connectors. Google says you can link more than 40 apps and services. The list includes Gmail, Google Drive, Google Calendar, Docs, Keep, YouTube, Maps, Spotify and WhatsApp, plus newer additions such as Canva and Dropbox.
You can also add a custom app through an MCP server URL. Google limits this to personal accounts held by users who are 18 or older and based in the US, in English, and you set it up in the web app. On work accounts, the list is shorter: the Google Workspace apps, Classroom and GitHub, switched on or off by your Workspace admin.
Developers have a third option. Gemini CLI reads MCP servers from an mcpServers block in its settings.json file.
Gemini Enterprise connectors and data stores
Connectors in Gemini Enterprise work through data stores: each connector you add creates one. For most third-party data sources, it creates one data store per entity type. Jira Cloud, for example, gets separate data stores for issues, attachments, comments and worklogs.
Connectors use one of two modes, and some let you choose:
- Federated search. Gemini queries the source system at question time and keeps no copy. There's no storage to manage, but Google notes that your query text is sent to every enabled third-party search backend. HubSpot, Supermetrics and Windsor.ai are examples of Gemini Enterprise federated connectors.
- Data ingestion. Gemini copies content into its own index. Search quality is usually better, but the copy needs storage and time to build.
Ingested data stores refresh on a schedule. A full sync replaces the data store's contents and catches deletions. An incremental sync adds new and updated items only, runs every 3 hours by default, and can be set anywhere from 3 hours to 7 days. Jira Cloud also offers real-time sync in preview.
Does Gemini have connectors like Claude?
Yes. Gemini connectors and Claude connectors both include a prebuilt catalog and support remote MCP servers. The main difference is control.
| Claude | Gemini app | Gemini Enterprise | |
|---|---|---|---|
| Who adds connectors | Individual users; owners on Team and Enterprise plans | Individual users (custom apps: US, 18+, personal accounts) | Admins only; users then sign in to approved connectors |
| Prebuilt catalog | Connectors directory | 40+ Connected Apps | 33+ (Business) or 138 (Standard and Plus) |
| Custom MCP | All plans; one custom connector on Free | Web app, eligible accounts | All editions, set up by an admin |
In Claude, a Pro user can add a custom connector from Customize > Connectors in a minute. In Gemini Enterprise, an admin enables each connector first, and users can only connect their own accounts to the ones that are approved. That makes setup slower, but it gives IT one place to see which enterprise data the company's AI agents can reach.
Gemini Enterprise connectors list by edition
Your edition decides how many Gemini Enterprise connectors you get. Google's connectors page lists 33 prebuilt connectors for Business and 138 for Standard and Plus. In October 2026, Google added 24 more to Business in public preview, including Gong, Intercom, Smartsheet and Zoho CRM. Business starts at $21 per seat per month, with a 30-day trial, for teams of up to 300. Standard and Plus are sold through sales, and launch coverage reported a starting price of $30 per seat per month on annual plans.
The table groups the Google Gemini Enterprise connectors by category. It names the main entries, not every one.
| Category | Business | Standard and Plus |
|---|---|---|
| Google Workspace | Gmail, Google Drive, Google Calendar, Google Chat, Google Sites | Same five |
| Google Cloud data | Not included | BigQuery, Cloud Storage, Cloud SQL, AlloyDB, Spanner, Firestore, Bigtable |
| Files and documents | Box, Dropbox, Microsoft OneDrive, SharePoint, Notion, Docusign, Smartsheet (preview) | Adds Egnyte, Airtable |
| Projects and engineering | Jira Cloud, Confluence Cloud, Asana, Monday, Linear, GitHub | Adds GitLab, Wrike, PagerDuty |
| Communication | Slack, Microsoft Teams, Microsoft Outlook | Adds Webex Meetings, Superhuman Mail |
| Sales and support | HubSpot, Zendesk; Zoho CRM, Intercom, Gong (preview) | Adds Salesforce, ServiceNow, Microsoft Dynamics 365 |
| Commerce and finance | Wix, Zoho Books (preview) | Adds Shopify, Stripe, Squarespace, Oracle NetSuite, Ramp |
| Marketing and analytics data | Not included | Supermetrics, Windsor.ai, Hex, Starburst, Nexla |
| Other | Excalidraw, Mermaid Chart, Hugging Face; Tavily, Gamma, Viator (preview) | Adds Microsoft Entra ID |
| Custom | Custom MCP server (preview) | Custom MCP server, custom connector |
Source: Google Cloud, Connectors for Gemini Enterprise editions, which has the full Gemini Enterprise connectors list. Checked October 2026.
What marketing teams won't find natively
The catalog has no connector for Google Ads, GA4, Meta Ads, LinkedIn Ads, TikTok Ads or Looker. Within the catalog, ad data reaches Gemini through two third-party Gemini data connectors, Supermetrics and Windsor.ai, both available only on Standard and Plus and both in public preview. On the Business edition, a custom MCP server is the only way to bring ad platform data in.
Gemini Enterprise third-party connectors: three ways in
Third-party data sources reach Gemini Enterprise by one of three routes. They differ in edition support and in how much setup your team owns.
- Prebuilt partner connectors. You pick the app from the catalog in the Google Cloud console and sign in. The partner maintains the API connection. Many partner connectors, including the marketing ones, are still in public preview, which means limited support from Google.
- Custom connectors. Google lists Gemini custom connectors only for Standard and Plus. They're meant for sources outside the catalog, such as internal systems. You prepare the data, map user identities and set access controls, so expect developer time.
- Custom MCP servers. Available on every edition, in public preview on Business. You register an MCP server URL as a data store. Google requires HTTPS with a certificate from a public authority and supports the Streamable HTTP transport only. Authentication is OAuth 2.0 or none. Actions are off by default, and you can enable up to 100 per server.
The third route covers most Gemini MCP connectors from data vendors. A vendor hosts the server and keeps the source APIs working, so your admin only fills in the connection details. You can also build and host the server yourself, which is covered in the reviews below.
Best Gemini data connectors compared
The table puts native Gemini Enterprise data connectors and third-party options side by side. Prices are entry plans with annual billing, checked on vendor sites in October 2026. Gemini Enterprise seats are billed separately by Google.
| Tool | Route into Gemini | Sources | Data freshness | Free option | Starting price |
|---|---|---|---|---|---|
| Coupler.io | Custom MCP server (all editions); Gemini CLI | 400+ | Scheduled: daily to every 15 min, by plan | Free plan (manual refresh); 7-day trial | $24/mo billed annually ($32 monthly) |
| Supermetrics | Prebuilt connector, Standard and Plus (preview) | 170+ | Fetched at question time | 14-day trial; Gemini access on request | Gemini pricing not published; plans from $44/mo |
| Windsor.ai | Prebuilt connector, Standard and Plus (preview) | 350+ | Fetched at question time | Free plan (1 source); 30-day trial | $19/mo billed annually ($23 monthly) |
| Databox | MCP for Gemini CLI | 130+ | Set by Databox plan | Free plan (no MCP); 14-day trial | $71/mo |
| BigQuery | Native, Standard and Plus | Your warehouse tables | Federated query (preview) or scheduled ingestion | Included; BigQuery usage billed | Included |
| Salesforce and HubSpot | Native (HubSpot on all editions, Salesforce on Standard and Plus) | CRM records | Federated search | Included | Included |
| Your own MCP server | Custom MCP server (all editions) | Any API you code against | Your choice | Open-source MCP SDKs | Developer and hosting time |
A closer look at each Gemini data connector
The reviews below cover each of these Gemini data connectors the same way: how it reaches Gemini, what it costs, where it fits, and what to check before you commit.
Coupler.io
Coupler.io brings ad, analytics and CRM data into Gemini as one blended dataset, so your team can compare spend and pipeline in a single question. It covers 400+ sources, including Google Ads, Meta Ads, GA4, HubSpot and Salesforce. You set up a data flow in Coupler.io, then register it in Gemini Enterprise as a custom MCP server with the endpoint and token it provides.
Before data reaches Gemini, you can join, append, filter and aggregate sources, or add formula columns. Coupler.io's Analytical Engine runs the calculations and aggregations, so Gemini receives computed figures to explain. The same data flows can also feed Claude, ChatGPT, Gemini CLI, Perplexity and Microsoft Copilot Studio.
- Pricing
- Billed by connected source accounts. Starter is $24/mo billed annually ($32 monthly) with daily refresh. Pro adds hourly refresh, and Custom plans refresh every 15 minutes. The free plan covers one source account with manual refresh.
Good fit for: teams that want blended ads, analytics and CRM data in one dataset, on any Gemini Enterprise edition.
Supermetrics
Supermetrics is a marketing data platform with 170+ sources. It sits in Google's catalog as a prebuilt connector for Standard and Plus and uses federated search in public preview, so Gemini fetches figures at question time.
The connector can also act on campaigns: Gemini can create, pause and update them across Google Ads, Meta, Microsoft Ads, TikTok and LinkedIn.
- Pricing
- The Gemini connector has no public price; access starts with a trial request or demo. Standard Supermetrics plans start at $44/mo annual billing ($55 monthly) for 3 sources.
Good fit for: ad teams on Standard or Plus who want to query and adjust campaigns from one chat.
Windsor.ai
Windsor.ai is a marketing data connector with 350+ sources, including Google Ads, Facebook, Instagram, TikTok, YouTube, Amazon Ads, LinkedIn and Shopify. It's a prebuilt Gemini Enterprise connector for Standard and Plus, in public preview. Gemini sends each question to Windsor.ai and stores nothing.
Write actions, such as pausing a campaign or changing a budget, are available too. You switch them on in Windsor.ai, and Gemini asks for confirmation before each one.
- Pricing
- Basic is $19/mo annual billing for 3 sources with daily refresh. Professional ($499/month on annual billing) refreshes every 15 minutes. A free plan covers one source, after a 30-day trial.
Good fit for: smaller teams that want ad data in Gemini at a low entry price.
Databox
Databox is a dashboard and reporting tool with 130+ integrations. Its MCP server works with Gemini CLI, along with Claude, ChatGPT, Cursor and n8n. Databox doesn't document a Gemini Enterprise setup.
- Pricing
- The Analyst plan starts at $71/month on annual billing. The free plan doesn't include MCP access.
Good fit for: analysts who already track KPIs in Databox and work with Gemini from the command line.
BigQuery (native)
The native BigQuery connector comes with Standard and Plus. Federated query, which Google recommends and is in preview, reads tables in place with no egress cost and respects BigQuery IAM and row-level security. Ingestion copies a table into a structured data store, once or on a schedule.
- Pricing
- Included in the edition; BigQuery query and storage costs still apply.
Good fit for: teams whose marketing data already lands in a warehouse.
Salesforce and HubSpot (native)
HubSpot is available on every edition and uses federated search, with a Manage CRM Objects action. A HubSpot admin authorizes the connector before users connect their own accounts. Salesforce comes with Standard and Plus; federated search is in public preview, and ingestion is in private preview with allowlisting.
- Pricing
- Included in the edition.
Good fit for: questions about deals, contacts and pipeline.
Your own custom MCP server
You can write an MCP server for any API with an open-source MCP SDK and host it yourself, for example on Cloud Run. Gemini Enterprise needs an HTTPS URL with a public certificate and the Streamable HTTP transport.
- Pricing
- No license fee; you pay for hosting and developer time.
Good fit for: internal systems or niche tools no vendor covers.
How to add connectors to Gemini
On Standard and Plus, Gemini Enterprise connectors are set up in the Google Cloud console by someone with the Discovery Engine Editor role. On Business, an admin turns connectors on under Manage team > Connected apps, and users then authorize their own accounts.
A prebuilt connector (Standard and Plus):
- In the Google Cloud console, open Gemini Enterprise > Data stores > Create data store.
- Select the source, such as Jira Cloud or HubSpot, and sign in. Some apps need their own admin to approve the connector first.
- Choose the entities to include and, where offered, federated search or data ingestion. Turn on any actions you want users to have.
- Pick a location (global, US or EU), name the data store and click Create. It's ready when the status shows Active.
A custom MCP server (Standard and Plus):
- Ask your Google Cloud admin to allow custom MCP data stores in the org policy and to add the server's domain to the allowed list.
- Go to Data stores > Create data store, search for Custom MCP Server and click Add MCP server.
- Choose OAuth 2.0 or no authentication. For OAuth, enter the server URL, authorization URL, token URL, scopes, client ID and client secret from your vendor, then click Verify Auth.
- Set the location and name, click Create, and wait for Active status.
- Open Actions > Reload custom actions and enable the tools you need, up to 100.
On Business, an admin adds a custom MCP server under Manage team > Connected apps > Add MCP Server. This option is in public preview.
In the personal Gemini app, Gemini connectors are switched on under Customize > Connectors. Custom MCP apps are added in the web app under Settings > Connected Apps > Add a custom app, with Keep Activity turned on.
FAQ
Are Gemini connectors free?
Connected Apps in the personal Gemini app cost nothing beyond your Google account, though some linked services have their own subscriptions. Gemini Enterprise connectors come with your edition's seat price. Third-party tools like Coupler.io, Supermetrics and Windsor.ai charge their own fees on top.
Can Gemini write data back to connected apps?
Some Gemini connectors support actions. Jira Cloud can create and update issues, Microsoft Outlook can send and forward mail, and HubSpot can manage CRM objects. On custom MCP servers, all actions start switched off until they're enabled in the data store settings.
Do I need Gemini Enterprise to query business data?
For company-wide access to CRM, ads and project data, yes. The personal Google Gemini app connects to Google services and a short list of others. Developers can also reach data sources through MCP servers in Gemini CLI without an Enterprise license.
Picking the right Gemini connector
The right Gemini Enterprise data connectors depend on your edition and where your data lives. Many teams combine two kinds of Gemini connectors: native ones for documents and a third-party tool for marketing data.
- Documents, tickets and chat: the native catalog covers Google Workspace, Microsoft OneDrive, Microsoft Teams, Jira Cloud, Confluence Cloud, Slack, Notion, Linear, Zendesk, Monday and PagerDuty.
- Ads, analytics and CRM in one place: Coupler.io works on every edition and blends sources before Gemini sees them.
- Ad data with campaign actions, on Standard or Plus: Supermetrics or Windsor.ai, with the preview status in mind.
- A warehouse already in place: the native BigQuery connector.
- Internal systems: a custom connector or your own MCP server.