Skip to content
  • There are no suggestions because the search field is empty.

BLAZE Integrations | How to Restrict BLAZE MCP Permissions & Store Access

BLAZE Model Context Protocol (MCP) strictly enforces the permissions of the signed-in BLAZE user account. You can easily restrict your AI assistant to read-only mode or limit its access to specific store locations.

Restricting MCP to Specific Stores

The AI assistant can only view and modify data for stores that the signed-in user is authorized to access.

To limit MCP access to a specific store location:

  1. Adjust User Permissions: Verify that the BLAZE user account used for MCP integration is assigned access only to the intended store locations.
  2. Use a Dedicated User Account: If an employee requires broad access within BLAZE but should have limited access when using MCP, create a separate BLAZE user profile assigned strictly to the allowed store(s) and sign in to MCP with those credentials.
  3. Authenticate MCP: Log into the BLAZE MCP connector using the configured user account.

TIP: If a user has access to multiple stores, specify the target store directly in your prompt (e.g., "Show sales at the Oakland store"). If omitted, the assistant will ask for clarification.

Restricting MCP to Read-Only Mode

You can enforce read-only access using three complementary safeguards:

1. User Role Permissions
The AI assistant cannot perform actions that the signed-in BLAZE user lacks permission to execute. Connecting MCP with a user profile set to view-only access prevents the assistant from editing or deleting data.
2. AI Application Tool Controls

Most AI applications allow granular control over connected tools and capabilities:

  • Claude: Navigate to the BLAZE Connector Settings to manage permissions. You can allow, require manual approval for, or block individual tools. Disabling tools that create, update, or delete data enforces strict read-only access.
  • Other AI Apps: Check your AI app’s connector or integration settings. Enterprise and business plans frequently provide additional administrative permissions controls.
3. Built-In BLAZE Safeguards
  • Action Approvals: The AI assistant prompts for user approval before creating, updating, or deleting any data in BLAZE.
  • Account Control: MCP functionality remains completely inaccessible until explicitly turned on by an account admin.
  • Audit Logs: Every action performed through MCP is recorded in your BLAZE activity logs and tagged as via MCP.