anchorOverview

Webex provides MCP servers that enable AI agents to securely access various Webex functionalities through the Model Context Protocol. These servers act as a bridge between your AI client and Webex services, allowing AI agents to perform actions and retrieve information on your behalf.

anchorAuthentication

Webex MCP servers support two authorization methods:

1. Token-Based Authentication

Use a WCIT (Webex Client Identity Token) to authenticate with Webex MCP servers. The WCIT is issued with only the spark:mcp scope which will allow to connect to MCP server. Additional scopes required by specific tools are requested at runtime via MCP elicitation during too call.

Note: This method only works with clients that support MCP elicitation. If your client does not support elicitation, use OAuth 2.0 instead.

1. Navigate to the Generate WCIT Token page in the Webex Developer Portal
2. Enter an appropriate name for your token
3. Click Generate New Token
4. Copy the generated token and store it securely (you'll need this for configuration)

Important: Keep your token secure and never commit it to version control.

2. OAuth 2.0

OAuth 2.0 can be used by clients that do not support elicitation. This requires creating a Webex Integration with the appropriate scopes.

1. Navigate to the Create a New Integration page
2. Fill in the form:
   • Will this integration use a mobile SDK? No
   • Integration name: Give any name
   • Icon: Choose one of the default icons
   • App Hub Description: Give any description
   • Redirect URI(s): You'll get this from the OAuth configuration page in your client
   • Scopes: Select the scopes required by your MCP server along with spark:mcp. To find the required scopes, go to the Agentic Apps page, choose your server, and click Learn More to open the MCP Server product page in Developer Portal where the required scopes are listed.
3. Click Add Integration to create your integration
4. On the following screen note down the Client ID and Client Secret — you won't be able to read your Client Secret again once you leave this page

anchorConfigure Your AI Client

Choose your AI client below and follow the configuration guide:

• ChatGPT
• Claude Code
• Copilot Studio
• Gemini CLI
• Salesforce Agentforce

anchorTroubleshooting

Common Issues

Connection Failed

Problem: MCP server fails to connect

Solutions:

• Verify your WCIT token is valid and not expired
• Check that the endpoint URL is correct
• Ensure you have an active internet connection

Authentication Error

Problem: "401 Unauthorized" or authentication errors

Solutions:

• Regenerate your WCIT token from the Developer Portal
• Ensure the token is properly formatted with "Bearer " prefix in the configuration
• Check that the token hasn't been revoked

Server Not Responding

Problem: MCP server is not responding to requests

Solutions:

• Verify your firewall isn't blocking the connection
• Try restarting your AI client

Tools Not Appearing

Problem: Webex tools don't show up in the AI client

Solutions:

• Confirm the configuration file syntax is correct (valid JSON)
• Restart the AI client completely
• Check the AI client's console/logs for error messages
• Verify you're using the latest version of your AI client:
  • Claude Desktop: Check for updates via Claude menu > "Check for Updates..."
  • VS Code: Ensure GitHub Copilot extension is up to date
  • Cursor: Check for IDE updates in Cursor settings

Getting Help

If you continue to experience issues, visit the Webex Developer Support.

anchorManaging Your MCP Connection

Unlinking an MCP Server

If you need to remove an MCP server from your AI client:

Claude Desktop:

1. Open Settings > Developer tab
2. Click "Edit Config" to open claude_desktop_config.json
3. Remove the server entry from the mcpServers object
4. Save the file and restart Claude Desktop

VS Code:

1. Use Command Palette: MCP: Remove Server and select the server to remove, or
2. Manually edit mcp.json (global or workspace) and remove the server entry from the servers object
3. Reload the VS Code window

Cursor IDE:

1. Open Cursor Settings > Tools & Integrations > MCP Tools
2. Remove the server entry from ~/.cursor/mcp.json
3. Save the file (Cursor will automatically detect the change)

Revoking Your WCIT Token

If your token is compromised or you no longer need it:

1. Navigate to the Manage WCIT Tokens page in the Webex Developer Portal
2. Locate the token you want to revoke in the list
3. Click the Delete button next to the token
4. Confirm the revocation

Important: After revoking a token, any AI clients using that token will immediately lose access to Webex MCP servers. You'll need to generate a new token and update your client configurations.

anchorNext Steps

Now that you've connected your Webex MCP server, you can:

• Explore the available Webex capabilities in your AI client
• Build custom workflows using Webex messaging and collaboration features
• Create intelligent agents that interact with your Webex organization

anchorAdditional Resources

• Webex Developer Portal
• Model Context Protocol Specification

Last Updated: February 2, 2026

Questions or Feedback? Visit the Webex Developer Community or contact support through the Developer Portal.