Setting Up Zephyr Local MCP
Zephyr MCP is available through SmartBear MCP Server and enables MCP-compatible AI clients to securely interact with Zephyr data and workflows.
Prerequisites
Before you begin, ensure that you have:
A Zephyr license.
A valid Zephyr API token.
Node.js 20 or later.
An MCP-compatible client, such as VS Code with GitHub Copilot, Claude Desktop, or Cursor.
Generate a Zephyr API token
Generate Zephyr API Token by following the instructions in the Zephyr API documentation.
Configure environment variables
Required variable
export ZEPHYR_API_TOKEN="your-zephyr-api-token"
Optional variable
Configure the API base URL based on your Jira instance's region.
export ZEPHYR_BASE_URL="https://api.zephyrscale.smartbear.com/v2"
Default value
https://api.zephyrscale.smartbear.com/v2
For regional endpoint details, see Zephyr APIs Regional Base URLs.
Install SmartBear MCP
Install the SmartBear MCP package globally:
npm install -g @smartbear/mcp
Or install it as a project dependency:
npm install @smartbear/mcp
Configure the MCP host
Configure your MCP-compatible client to use SmartBear MCP Server.
Supported client configurations include:
VS Code with GitHub Copilot
Claude Desktop
Cursor
Additional Configuration and Integrations
SmartBear MCP Server also supports integrations with other SmartBear products, including:
Swagger
Reflect
QMetry
BugSnag
For advanced configuration details, multi-product setup guidance, and complete installation instructions, see the SmartBear MCP documentation.