What Is MCP?
The Model Context Protocol (MCP) provides a standardized way for AI assistants and other compatible AI tools to securely connect to and access information from Crelate. With MCP, you can bring your Crelate recruiting data into supported AI tools such as ChatGPT, Claude, Microsoft Copilot, and other compatible solutions without building a custom integration.
Most Crelate MCP tools are read-only, allowing supported AI tools to retrieve information from Crelate without making changes to your records. A limited number of tools can create, update, or delete information in Crelate. Depending on the AI client you use, you may be prompted to approve these actions, or you may have the option to allow them automatically.
Connecting to Your AI Clients
How the Connection Works
You can connect Crelate by signing in with your Crelate credentials directly through your AI client. There is no key to generate, copy, or paste.
Once connected, the AI client acts on your behalf using the permissions associated with your Crelate user account. This means the AI client can only access the Crelate records and functionality that your account is already authorized to access.
Most Crelate MCP tools are read-only. For tools that can create, update, or delete information, the AI client may require you to approve the action before it is completed. Some AI clients, such as Claude, may also allow you to configure these tools to Always allow instead.
Before You Begin
You will need the following:
The Crelate MCP server address:
https://mcp.crelate.com/mcpAn AI client that supports remote MCP connectors with sign-in, such as Claude.
Connector permissions in your AI client. If your organization uses a Claude Team or Enterprise plan, an organization owner may need to add the Crelate connector before individual users can connect.
MCP access on your subscription. The Crelate MCP Server is available with Business Plus and higher subscriptions, depending on the services included in your plan, and is enabled per organization by Crelate.
If MCP has not yet been enabled for your organization, contact your Success Manager or Crelate Support to review your subscription and confirm the next steps. If MCP is enabled and you do not see the AI / MCP Access option in Your Profile & Preferences, contact your Crelate administrator to confirm that the MCP Access role privilege has been enabled for your role.
Connect Crelate to Claude
In Claude, select Customize from the left sidebar.
Select Connectors, then select Add > Custom > Web.
In the Add custom connector window, enter a name for the connector, such as Crelate MCP. This is the name that will appear in your connectors list.
In the server address field, enter
https://mcp.crelate.com/mcp.
Review the connection settings. Claude should detect the appropriate options automatically, so in most cases, you will not need to change them:
Authentication: Leave this set to Always required. Each user will sign in with their own Crelate account before using the connector.
OAuth client: Leave this set to Use Anthropic's hosted client metadata, which is the recommended option.
Save the connector and select Connect. A Crelate sign-in window will open.
Sign in with your Crelate credentials and approve the requested access.
Once the connection is complete, the Crelate connector will appear under Customize > Connectors in Claude.
You can now ask Claude questions about your Crelate data, and Claude will use the Crelate connector to retrieve the information based on your existing Crelate permissions.
Connect Crelate to Microsoft 365 Copilot
Microsoft 365 Copilot connects differently from the sign-in-only clients described in this article. A Microsoft 365 administrator completes a one-time setup, and then each user signs in to Crelate from Copilot. The scopes in this guide give Copilot read access to your Crelate data, and Copilot can only see what your Crelate user is already allowed to see.
Quick Reference
Who completes the setup: A Microsoft 365 administrator with the AI Administrator or Global Administrator role completes Steps 1 through 4. Each user then completes Steps 5 and 6.
Values you will need
Item | Value |
Crelate MCP server address |
|
Scopes |
|
Authorization and token endpoints | Listed at |
Redirect URI, Client ID, and Client secret | Generated for you in Crelate (Step 1) |
Steps at a glance
Generate credentials in Crelate. In the AI Platform Connectors area in settings, generate credentials for Microsoft 365 Copilot. Save the Client secret, which is shown only once.
Create the app. In the Microsoft Developer Portal (
dev.teams.microsoft.com), create an app named Crelate MCP and copy its App ID.Register the OAuth client. In the Developer Portal, go to Tools, then OAuth client registration. Enter the Client ID, Client secret, endpoints, and scopes, enable PKCE, and keep the default client authentication method. Copy the Reference ID.
Create the connector. In the Microsoft 365 admin center, go to Copilot, then Connectors, and add a Custom MCP connector. Enter the server address and App ID, choose OAuth 2.0, enter the Reference ID, and select Authorize.
Connect as a user. In Copilot, open Chat settings, then Sources, and select Connect on Crelate MCP. Sign in to Crelate and allow access.
Start chatting. Type @ in a new chat and choose Crelate MCP, then ask a question such as “List my open Jobs in Crelate.”
For screenshots, field-by-field instructions, and troubleshooting, see Connecting Microsoft 365 Copilot to the Crelate MCP Server below.
Connect Crelate to Other AI Clients
The setup steps vary by AI client, but the process generally follows the same pattern:
Open your AI client's Connector or MCP server settings.
Add a new remote MCP server using the address
https://mcp.crelate.com/mcp.When prompted, sign in with your Crelate credentials and approve the requested access.
For client-specific setup instructions, refer to your AI client's documentation.
View and Manage Connected AI Clients
You can review the AI clients connected to your Crelate account and revoke access at any time.
In Crelate, navigate to Your Profile & Preferences.
Select AI / MCP Access.
Before you connect an AI client, the Connected apps section will show that no clients are connected.
After you connect an AI client, it will appear in the Connected apps list along with the access it has been granted and the date and time it was connected.
Select the scopes link under Access to view the permissions granted to the connected AI client:
Scope | What it allows |
read | View individual Crelate records |
read.bulk | View multiple records at once |
write | Create and update individual records |
write.bulk | Create and update multiple records at once |
delete | Remove records |
To remove an AI client's access, select Disconnect. Access is revoked immediately. You can reconnect the AI client at any time by repeating the setup steps in your AI client.
Note: Only AI clients that connect by signing in appear in the Connected apps list. Personal MCP access keys created before sign-in connections were available are managed separately.
Manage Existing MCP Access Keys
If you created a personal MCP access key before Crelate introduced sign-in connections, you can continue to view and manage that key from AI / MCP Access.
Existing MCP access keys can still be revoked, but new MCP access keys can no longer be created. To connect a new AI client, use the sign-in method described above.
Important: If you have an existing MCP access key, keep it secure and do not share it with others. Anyone with access to the key may be able to access Crelate data using the permissions associated with your account.