How to connect Xobin MCP to Gemini Enterprise?
Unlike Claude and ChatGPT, Gemini Enterprise does not auto-register an OAuth client. An administrator must register Gemini as an OAuth app on Xobin, then create a Custom MCP Server data store.
Note: **Gemini only supports Streamable HTTP (the Xobin URL ending in /mcp). Legacy SSE is not supported. Before proceeding, make sure you have reviewed the prerequisites and connection values in** Connecting Xobin to AI Assistants via MCP.
Step 1: Register Gemini as an OAuth Client on Xobin
- Request a pre-registered Gemini client from Xobin, or register one directly against the MCP authorization server.
- Gemini requires the following redirect URI:
https://vertexaisearch.cloud.google.com/oauth-redirectExample registration request:
curl -sS -X POST "https://{MCP_HOST}/register" \
-H "Content-Type: application/json" \
-d '{
"client_name": "Gemini Enterprise",
"redirect_uris": [
"https://vertexaisearch.cloud.google.com/oauth-redirect"
],
"grant_types": ["authorization_code", "refresh_token"],
"response_types": ["code"],
"token_endpoint_auth_method": "none",
"scope": "mcp:read mcp:write"
}'
Save the returned client_id (and client_secret, if provided) in a secure secrets store — never in chat logs.
Step 2: Google Cloud Gemini Enterprise
Your Google Cloud project must allow Custom MCP data stores and egress to {MCP_HOST}. The user creating the data store needs the Discovery Engine Editor role (roles/discoveryengine.editor).
- In the Google Cloud console, open Gemini Enterprise.
- Go to Data Stores → Create Data Store.
- Search for Custom MCP Server and click Add MCP Server.
- Select OAuth 2.0 and enter the following:
Field | Value |
|---|---|
MCP Server URL | |
Authorization URL | |
Token URL | |
Authorization URL Parameters |
|
Client ID | From Step 1 |
Client Secret | From Step 1, or leave blank if the client is public |
Enable PKCE Support | On |
Scopes |
|
client_id, redirect_uri, and scope. The resource parameter tells Xobin which MCP resource to issue tokens for.- Click Verify Auth, sign in to Xobin, and approve access.
- Choose a location, name the data store Xobin, and click Create.
- Wait until the data store status shows Active.
- Open Actions, reload custom actions, and enable the required Xobin tools (up to 100).
- Connect the data store to your Gemini Enterprise app. In the chat UI, authorize the connector if it still shows Authorize, then enable it.
Step 3: Gemini Enterprise – Business Edition
- Sign in at business.gemini.google.
- Go to Settings & Help.
- Select your team → Manage Team → Connected Apps.
- Click Add MCP Server.
- Enter the name Xobin, a short description of the hiring workflow, and the MCP Server URL https://mcp.xobin.com/mcp.
- Fill in the same Authorization URL, Token URL, Client ID, Client Secret, PKCE, scopes, and
&resource=https://{MCP_HOST}/mcpparameter as above. - Click Add, then enable the connection so your team can access it.
- In chat, open Connectors, authorize Xobin, and keep it toggled on.
Confirming the Connection:
- Test the connection with: "Using Xobin, what job tracks are open, and how many candidates are in each?"
- Then test a write action you can safely review: "Draft (do not send) an assessment invitation for a candidate I name."
Troubleshooting:
- Gemini "Verify Auth" fails.
Check the MCP URL (/mcp), Authorization URL (/authorize with no query string), Token URL (/token), PKCE settings, scopes (mcp:read mcp:write), and the resource parameter. Ensure the MCP host is allowed as an egress FQDN in your Google Cloud organization policy.
- OAuth redirect error ("redirect_uri is not registered").
The registered redirect URI must exactly match https://vertexaisearch.cloud.google.com/oauth-redirect.
Updated on: 08/14/2026
Thank you!
