Connect your AI tool to Yeeflow through one remote MCP server, sign in with OAuth, and work with the Yeeflow features your account is permitted to use.
Yeeflow provides a unified Model Context Protocol (MCP) service. A single connection brings together business operations, application building, organization administration, and Service Portal management. You no longer need to configure a separate server for each capability area.
Connection details
Suggested connection name: Yeeflow
Server URL: https://api.yeeflow.com/v1/mcp
Connection type: Remote MCP over HTTPS
Authentication: OAuth
What you can do
The unified service includes the following capability areas. The actions you can complete depend on your Yeeflow permissions and the tools supported by your AI integration.
Operations: Work with business records, files, workflow tasks, agents, and webhooks.
App Builder: Build and manage applications, workspaces, application resources, packages, and portals.
Admin: Manage organization users, departments, groups, locations, and positions.
Service Portal: Manage Service Portal users, groups, and related access settings.
These are capability areas within the same connection. You do not need a separate MCP connection for each area.
Before you start
Use an AI tool that supports remote MCP servers over HTTPS and OAuth authentication.
Make sure your AI tool or organization allows you to add a remote MCP connection.
Have a Yeeflow account with access to the organization, applications, and features you intend to use.
Ask your Yeeflow administrator for any additional product permissions required for your work.
Connecting through MCP does not grant additional Yeeflow permissions.
Add the Yeeflow connection
Open your AI tool’s MCP, Connectors, Integrations, or Developer settings. The exact labels depend on your AI tool.
Choose the option to add a remote MCP server.
Enter Yeeflow as the connection name.
Enter https://api.yeeflow.com/v1/mcp as the server URL.
Select OAuth if prompted for an authentication method.
Complete the Yeeflow sign-in and consent flow in your browser. Use the account you intend to work with.
Return to your AI tool and confirm that the connection is available. Refresh its tool list or start a new conversation if needed.
Do not paste access tokens, API keys, passwords, cookies, or authorization headers into the MCP configuration. Authentication is handled through OAuth.
If your AI tool already provides a Yeeflow integration, use its connection flow and check its setup instructions before adding a duplicate manual connection.
Verify your connection
Start with a read-only request: Show which Yeeflow account I am connected as.
Confirm that the returned account is the one you intended to use. Then try a request relevant to your role, such as Show my to-do workflow tasks. If you have the relevant workspace access, you can also try List my Yeeflow workspaces.
An access-denied response for a particular action does not necessarily mean the connection failed. The action may require permissions your account does not have.
Update an existing connection
If you previously configured separate App Builder, Operations, Admin, or Service Portal MCP connections, move to the unified connection. The previous configuration used URLs ending in /mcp/app-builder, /mcp/operations, /mcp/admin, and /mcp/service-portal.
Add the unified connection using https://api.yeeflow.com/v1/mcp and complete OAuth sign-in.
Verify your account and test a read-only request relevant to your work.
Update any AI assistants or saved configurations that explicitly reference the previous connections. Refresh their available tools.
Once the unified connection works for your intended tasks, remove or disable the redundant connections in that AI tool.
For an integration managed by your organization, coordinate the configuration update with its administrator.
Permissions and safe use
Your existing Yeeflow permissions still apply through the unified service:
Business operations require access to the relevant records, applications, and workflow actions.
Application structure, settings, and resource creation require the appropriate application-administrator permissions. Editor or Visitor access does not provide application-administrator rights.
Organization administration and Service Portal management retain their own permission requirements.
Before creating, changing, deleting, installing, upgrading, submitting, or changing access, review the target and intended changes. Ask the AI tool to show the result after an action completes.
A successful configuration change does not by itself confirm that an application, workflow, or portal behaves correctly. Test the affected functionality in Yeeflow when appropriate.
Troubleshooting
Authentication fails or the connection expires: Reconnect the unified service and complete OAuth sign-in. Check that you are using an active Yeeflow account.
Connected with the wrong account: Disconnect or sign out through your AI tool, then reconnect with the intended account. Verify the connected identity before continuing.
Connected, but tools do not appear: Confirm the unified URL, refresh the tool list, or start a new conversation. If needed, re-add the connection and complete OAuth again. Check your AI tool’s remote MCP support and organization settings.
A particular action is denied: Ask your Yeeflow administrator to check the required permissions for the target application or resource. Adding another MCP connection does not grant those permissions.
A saved assistant still references an old connection: Update its connection selection to the unified service and refresh its tools.
Your AI tool does not support OAuth: Use an AI tool that supports remote HTTPS MCP and OAuth. Do not replace OAuth by pasting credentials into its configuration.
Next steps
Start with a read-only task relevant to your role. Once you have verified your account and access, use the same connection for business operations, application building, or administration as your permissions allow.