MCP Server

The Brie MCP server lets AI agents interact with your Brie bug reports via the Model Context Protocol.

Install

npm install -g @briehq/mcp

Or run directly:

npx -y @briehq/mcp

Setup wizard

The fastest way to get started:

briehq-mcp init

This walks you through entering your API key, testing the connection, and generating the config for your AI client.

Claude Desktop

Add to your claude_desktop_config.json:

{
  "mcpServers": {
    "brie": {
      "command": "npx",
      "args": ["-y", "@briehq/mcp"],
      "env": {
        "BRIE_API_KEY": "brie_pat_your_key_here"
      }
    }
  }
}

Claude Code

claude mcp add brie -e BRIE_API_KEY=brie_pat_your_key_here -- npx -y @briehq/mcp

Available tools

Tool	Scope	Description
`list_slices`	`slices:read`	List bug reports with filtering and pagination
`get_slice`	`slices:read`	Get a single bug report by ID
`create_slice`	`slices:write`	Create a new bug report
`update_slice`	`slices:write`	Update status, priority, description, labels
`delete_slice`	`slices:delete`	Delete a bug report
`list_workspaces`	`workspaces:read`	List accessible workspaces
`get_workspace`	`workspaces:read`	Get workspace details
`list_spaces`	`spaces:read`	List spaces within a workspace
`get_me`	`user:read`	Get current user profile

Example prompts

Once connected, try asking your AI agent:

"List all open bugs in my workspace"
"Create a bug report: checkout page throws 500 on mobile Safari"
"Mark slice abc123 as resolved"
"What workspaces do I have access to?"

Configuration

Variable	Required	Default	Description
`BRIE_API_KEY`	Yes	—	Your Brie API key
`BRIE_API_URL`	No	`https://api.briehq.com`	API base URL (use for local dev)