Connect Agentforce to TaskRay MCP

This guide connects Salesforce Agentforce to your TaskRay data. Agentforce has a native MCP client, so a Salesforce admin or maker registers TaskRay MCP as an external MCP server, adds its tools to an agent, and each user signs in with their own Salesforce login the first time they use it. Every connection stays scoped to that user's own permissions.

This is not the prebuilt TaskRay AI agent This guide is about connecting your own Agentforce agent to TaskRay MCP. TaskRay also offers TaskRay AI, a prebuilt Agentforce agent that ships as its own managed package with its own setup. That is a separate product and does not use the connector described here. You can use either or both.
Before you start You'll need a TaskRay MCP beta enrollment for your org (your admin handles this, see the Administrator Installation & Setup guide), both permission sets assigned in Salesforce to every user who will use the agent (TaskRay Agents App Access and TaskRay Agents - User), and Agentforce set up in your org with permission to register an MCP server and build an agent.

Admin or maker: connect TaskRay MCP to Agentforce

  1. Get your org's connector URL. Open the TaskRay connection page and sign in once with Salesforce. Use https://mcp.taskray.com/oauth/login for a production org, or https://mcp.taskray.com/oauth/login?sandbox=true for a sandbox or scratch org. Copy the connector URL shown on the connected page.
  2. In Salesforce Setup, register TaskRay MCP as an external MCP server in Agentforce, using that connector URL as the endpoint and OAuth as the authentication method so each user signs in as themselves.
  3. Use the tool allowlist to choose which TaskRay tools the agent can use.
  4. Add the TaskRay tools to your agent in Agent Builder, then activate the agent.
Registering an MCP server in Agentforce Registering an external MCP server and adding its tools to an agent are Salesforce steps that live in Setup and Agent Builder. For the exact navigation and options, see Salesforce's Agentforce MCP.

End users: sign in on first use

Once the agent is active, each user signs in with their own Salesforce credentials the first time they use a TaskRay action. There is nothing for an end user to paste. They only complete the Salesforce sign-in prompt, and the agent then works with TaskRay data on their behalf.

Every user must be provisioned in Salesforce Registering the MCP server does not grant anyone data access. Each user still needs both permission sets in Salesforce, or their sign-in fails with "access denied." Access always mirrors that user's own Salesforce permissions, and there is no shared service account.
That's it! Your Agentforce agent can now work with TaskRay data on each user's behalf. See the First Prompts to Try after Connecting TaskRay MCP for what to ask.

Frequently Asked Questions

Q. How is this different from TaskRay AI?
TaskRay AI is a prebuilt Agentforce agent that TaskRay ships as its own managed package, with its own setup, and it does not use MCP. This guide instead connects your own Agentforce agent to TaskRay MCP through the connector URL. You can use either or both.

Q. A user gets "access denied" the first time they use the agent.
They aren't provisioned in Salesforce yet. Assign both permission sets (TaskRay Agents App Access and TaskRay Agents - User) and confirm App Access is selected on the External Client App.

Q. Sign-in works, but TaskRay actions fail.
The user is missing the TaskRay Agents - User permission set, which grants the API Enabled entitlement every tool call needs.

Q. Should I use the production or sandbox connector URL?
Match your target org. Use ?sandbox=true for sandbox and scratch orgs so sign-in goes to test.salesforce.com.

Still stuck? Contact TaskRay Support with the exact error text and which org (production or sandbox) you're connecting.

Was this article helpful?

0 out of 0 found this helpful

Have more questions? Submit a request