Getting started with MCP

• MCP server URL:

  • Staging: https://app.staging.360learning-dev.com/api/mcp
  • Production: https://app.360learning.com/api/mcp

• Client ID: 360learning-mcp

• MCP server URL:
  • Staging: https://app.us.staging.360learning-dev.com/api/mcp
  • Production: https://app.us.360learning.com/api/mcp
• Client ID: 360learning-mcp

Steps may change over time. For the latest instructions, see Claude's official connector documentation.

If you're on a Team or Enterprise plan

An Owner or Primary Owner needs to do this once first:

1. Open Claude.
2. Go to Organization settings > Connectors.
3. Click Add, hover over Custom, then select Web.
4. Enter the MCP Server URL.
5. Click Advanced settings and enter the OAuth Client ID. Leave the Client secret field empty.
6. Click Add.

Then, each member connects individually:

1. Go to Customize > Connectors.
2. Find the connector your Owner added (usually labeled "Custom").
3. Click Connect to authenticate.
4. Log in with your 360Learning credentials and select a platform if prompted.
5. Once logged in, you'll be redirected back to Claude web/desktop.

If you're on a Pro or Max plan

1. Open Claude.
2. Go to Customize > Connectors.
3. Click Add, then select Add custom connector.
4. Give the connector a name.
5. Enter the MCP Server URL, then click Continue.
6. Select Use your own OAuth client, then enter the Client ID. Leave the Client secret field empty.
7. Click Add.
8. Click Connect to authenticate.
9. Log in with your 360Learning email and password and select a platform if prompted.
10. Once logged in, you'll be redirected back to Claude web/desktop.

Claude is now connected to the 360Learning MCP server.

Steps may change over time. For the latest instructions, see Claude Code official connector documentation.

1. Run the following command in your terminal (not inside a claude session), using the MCP server URL for your region and environment:

   claude mcp add --transport http --client-id 360learning-mcp --callback-port 8765 360learning <MCP_SERVER_URL>

   Example for EU staging: claude mcp add --transport http --client-id 360learning-mcp --callback-port 8765 360learning https://app.staging.360learning-dev.com/api/mcp

   ⚠️ Always use the 8765 port, as http://localhost:8765/callback and http://127.0.01:8765/callback are the only authorized redirect URIs for Claude Code.

2. Run /mcp inside a Claude Code session.

3. Select the 360Learning server, then choose Authenticate.

4. A browser tab opens to the 360Learning login page. Log in with your 360Learning email and password and select a platform if prompted.

5. Once logged in, you'll be redirected back to your terminal.

Claude Code is now connected to the 360Learning MCP server.

Steps may change over time. For the latest instructions, see OpenAI's official plugin documentation.

If you're on an Enterprise plan

A Workspace Admin needs to do this once first:

1. Open ChatGPT.
2. Go to Workspace settings > Plugins.
3. Turn plugins on for the workspace (off by default on Enterprise).
4. In Settings → Security or Settings → Apps & Connectors → Advanced, enable Developer mode.
5. On the Plugins page, click the + button, and enter the following information:
   1. In the Connection field, enter the MCP Server URL .
   2. Keep OAuth as Authentication.
   3. Go into Advanced OAuth settings.
   4. Select User-Defined OAuth Client as the Registration method.
   5. Use 360learning-mcp as the OAuth Client ID.
   6. Select none as the Token endpoint auth method.
   7. Tick the box I understand and want to continue.
   8. Click Create.
6. Click Workspace Settings → Apps (or Plugins).
7. Open Drafts.
8. Select your plugin.
9. Click Publish.
   The plugin is added to the workspace's approved connectors list.
10. Click Workspace Settings → Plugins to confirm plugins are turned on for the workspace (Enterprise starts with this off).
11. Set which users, groups, or roles can see and use the newly published custom plugin, using your organization's RBAC controls.

Then, each member connects individually:

1. Click Settings → Apps (or the Plugin Directory).
2. Select the plugin in the list.
3. Click Connect to authenticate.
4. Log in with your 360Learning credentials and select a platform if prompted.
5. Once logged in, you'll be redirected back to ChatGPT.

If you're on a Business plan

1. Open ChatGPT.
2. Click Settings.
3. Click Security and login.
4. Turn on Developer mode.
5. In the Plugins page, click the + button, and enter the following information:
   1. Enter the MCP Server URL in the Connection field.
   2. Keep OAuth as Authentication and go into Advanced OAuth settings.
   3. Select User-Defined OAuth Client as the Registration method.
   4. Use 360learning-mcp as the OAuth Client ID.
   5. Select none as the Token endpoint auth method.
   6. Tick the box I understand and want to continue.
   7. Click on Create.
6. Go to Settings → Apps (or the Plugin Directory).
7. Select the plugin in the list.
8. Click Connect to authenticate.
9. Log in with your 360Learning credentials and select a platform if prompted.
10. Click Connect to authenticate.
11. Once logged in, you'll be redirected back to ChatGPT.

ChatGPT is now connected to the 360Learning MCP server.

Steps may change over time. For the latest instructions, see Dust's official tools documentation.

A Dust Admin needs to do this once first:

1. Open Dust.
2. Go to Spaces > Tools.
3. Click + Add Tools, and then + Add MCP Server.
4. In the Configure MCP Server dialog box, enter the following information:
   1. In URL, enter the MCP Server URL for your region and environment.
   2. In Authentication, choose Static OAuth.
   3. In the How do you want to connect? section, keep Personal accounts. This ensures every MCP interaction is made on behalf of your users, inheriting from their rights and profiles on the 360Learning platform.
   4. In OAuth Client ID, enter 360learning-mcp .
   5. Leave OAuth Client Secret blank.
   6. In OAuth Token Endpoint and OAuth Authorization Endpoint, use the following URLs:
      • OAuth Token Endpoint:

        • Staging: https://public-api.cloud-iam.com/realms/mcp-staging-eu/protocol/openid-connect/token
        • Production (Available starting September 21, 2026): https://public-api.cloud-iam.com/realms/mcp-prod-eu/protocol/openid-connect/token

      • OAuth Authorization Endpoint:

        • Staging: https://public-api.cloud-iam.com/realms/mcp-staging-eu/protocol/openid-connect/auth
        • Production (Available starting September 21, 2026): https://public-api.cloud-iam.com/realms/mcp-prod-eu/protocol/openid-connect/auth
   7. In OAuth Scope(s), enter openid offline_access .
   8. In OAuth Resource / Audience, re-use the MCP Server URL for your environment and region.
   9. Leave Token Endpoint Authentication Method set to Request body.
   10. Do not toggle Use custom headers on.
   11. Click Setup connection.
5. Complete authentication on 360Learning.
6. Click Save.
7. In the right-side panel opening, click Save.

After that, the first time a member uses a 360Learning tool in Dust, they'll be prompted to sign in and authenticate with 360Learning.

Customer Success Partner (CSP) or Solutions Engineer (SE)Steps may change over time. For the latest instructions, see Microsoft Copilot's official tools documentation.

A Copilot Admin needs to do this once first:

1. Open Copilot Studio.
2. Go to Agents.
3. Select an existing agent or create a new one.
4. In the agent configuration page, click on Tools > + Add > Model Context Protocol (MCP).
5. Enter the following information:
   1. Choose a Server name and Server description.
   2. In Server URL, use the MCP Server URL for your region and environment.
   3. For Authentication, choose OAuth 2.0.
   4. For Configuration type, choose Manual.
   5. In Client ID, enter 360learning-mcp-confidential .
   6. Reach out to your Customer Success Partner (CSP) or Solutions Engineer (SE) to get the Client secret, specifying your region and environment.
   7. In Authorization URL, Token URL and Refresh token URL, use the following URLs:
      • Token URL & Refresh token URL:

        • Staging: https://public-api.cloud-iam.com/realms/mcp-staging-eu/protocol/openid-connect/token
        • Production: https://public-api.cloud-iam.com/realms/mcp-prod-eu/protocol/openid-connect/token

      • Authorization URL:

        • Staging: https://public-api.cloud-iam.com/realms/mcp-staging-eu/protocol/openid-connect/auth
        • Production: https://public-api.cloud-iam.com/realms/mcp-prod-eu/protocol/openid-connect/auth
   8. In Scopes, enter openid offline_access .
6. Click on Add.
7. In the Save your redirect URL dialog, click on Next as Copilot redirect URLs are already whitelisted on our end.
8. In the Select a connection dialog, click on Create new connection > Create.
9. Complete authentication on 360Learning.
10. Click Add.
11. Save and publish your agent.

After that, the first time a member uses your agent on Microsoft Copilot, they'll be prompted to sign in and authenticate with 360Learning.

Steps may change over time. For the latest instructions, see Snowflake's official tools documentation.

To configure the agent leveraging 360Learning's MCP server, prerequisites are:

• A SNOWFLAKE INTELLIGENCE object is created.
• A Snowflake user with a role having the following right: CREATE INTEGRATION , CREATE EXTERNAL MCP SERVER , CREATE AGENT and MODIFY right on the created SNOWFLAKE INTELLIGENCE object.

Then, a Snowflake user with a role having the above rights needs to:

1. Create the API integration

CREATE API INTEGRATION $API_INTEGRATION_NAME
  API_PROVIDER = external_mcp
  API_ALLOWED_PREFIXES = ('$MCP_SERVER_PREFIX')
  API_USER_AUTHENTICATION = (
    TYPE = OAUTH2
    OAUTH_CLIENT_ID = '$MCP_CLIENT_ID'
    OAUTH_CLIENT_SECRET = '$MCP_CLIENT_SECRET'
    OAUTH_CLIENT_AUTH_METHOD = CLIENT_SECRET_BASIC
    OAUTH_AUTHORIZATION_ENDPOINT = '$MCP_AUTH_URL'
    OAUTH_TOKEN_ENDPOINT = '$MCP_TOKEN_URL'
    OAUTH_ALLOWED_SCOPES = ($MCP_SCOPES)
    OAUTH_REFRESH_TOKEN_VALIDITY = 1209600
  )
  ENABLED = TRUE;

While interpolating the variables with the following information:

• $API_INTEGRATION_NAME : Any name you would like to use to name your API integration.
• $MCP_SERVER_PREFIX : Use the base URL (i.e. without the /api/mcp) of the MCP Server URL for your region and environment.
• $MCP_CLIENT_ID : Use 360learning-mcp-confidential .
• $MCP_CLIENT_SECRET : Reach out to your Customer Success Partner (CSP) or Solutions Engineer (SE) to get this value, specifying your region and environment.
• $MCP_AUTH_URL & $MCP_TOKEN_URL : use the following URLs
• • $MCP_TOKEN_URL:

    • Staging: https://public-api.cloud-iam.com/realms/mcp-staging-eu/protocol/openid-connect/token
    • Production: https://public-api.cloud-iam.com/realms/mcp-prod-eu/protocol/openid-connect/token

  • $MCP_AUTH_URL:

    • Staging: https://public-api.cloud-iam.com/realms/mcp-staging-eu/protocol/openid-connect/auth
    • Production: https://public-api.cloud-iam.com/realms/mcp-prod-eu/protocol/openid-connect/auth
• $MCP_SCOPES : Use 'openid', 'offline_access' .

2. Create the external MCP server

CREATE EXTERNAL MCP SERVER $DB_NAME.$SCHEMA_NAME.$MCP_SERVER_NAME
WITH DISPLAY_NAME = '$MCP_DISPLAY_NAME'
URL = '$MCP_SERVER_URL'
API_INTEGRATION = $API_INTEGRATION_NAME;

While interpolating the variables with the following information:

• $DB_NAME.$SCHEMA_NAME.$MCP_SERVER_NAME : Database name, schema name and MCP server name of your choice.
• $MCP_DISPLAY_NAME : Display (human-friendly) name for your MCP server.
• $API_INTEGRATION_NAME : Name of the API integration you've created above.

Then, the same Snowflake user needs to:

3. Go into Agent Studio.
4. Click on Create agent.
5. Select the database and schema on which you've created the MCP server at step 2.
6. Choose an Agent object name and Display name.
7. Click on Create agent.
8. In the agent page, go to the Configuration tab > MCP tab.
9. Find the external MCP server created at step 2 and click on Add to agent.
10. Save and Publish your agent.

After that, the first time a member uses your agent on Snowflake, they'll be prompted to sign in and authenticate with 360Learning.

To connect other AI assistants, 360Learning must first whitelist the assistant's callback URL. Contact your Customer Success Partner (CSP) or Solutions Engineer (SE) and provide the assistant's exact callback URL to request whitelisting.

Once whitelisted, add the 360Learning MCP server as a custom connector or remote MCP server in your assistant, using the values above. Steps vary by assistant. See its own documentation for adding a remote MCP server.