smartbear
SmartBear's official MCP Server
claude mcp add --transport stdio smartbear-smartbear-mcp npx -y @smartbear/mcp@latest \
--env QMETRY_API_KEY="${input:qmetry_api_key}" \
--env QMETRY_BASE_URL="${input:qmetry_base_url}" \
--env SWAGGER_API_KEY="${input:swagger_api_key}" \
--env ZEPHYR_BASE_URL="${input:zephyr_base_url}" \
--env ZEPHYR_API_TOKEN="${input:zephyr_api_token}" \
--env PACT_BROKER_TOKEN="${input:pact_broker_token}" \
--env REFLECT_API_TOKEN="${input:reflect_api_token}" \
--env BUGSNAG_AUTH_TOKEN="${input:bugsnag_auth_token}" \
--env PACT_BROKER_BASE_URL="${input:pact_broker_base_url}" \
--env PACT_BROKER_PASSWORD="${input:pact_broker_password}" \
--env PACT_BROKER_USERNAME="${input:pact_broker_username}" \
--env SWAGGER_UI_BASE_PATH="${input:swagger_ui_base_path}" \
--env COLLABORATOR_BASE_URL="${input:collab_base_url}" \
--env COLLABORATOR_USERNAME="${input:collab_username}" \
--env BUGSNAG_PROJECT_API_KEY="${input:bugsnag_project_api_key}" \
--env SWAGGER_PORTAL_BASE_PATH="${input:swagger_portal_base_path}" \
--env COLLABORATOR_LOGIN_TICKET="${input:collab_login_ticket}" \
--env SWAGGER_REGISTRY_BASE_PATH="${input:swagger_registry_base_path}"How to use
SmartBear MCP server exposes SmartBear APIs through an MCP interface, enabling AI assistants to securely access BugSnag, Reflect, Swagger, Pact-related tools, QMetry, Zephyr, and Collaborator from natural language prompts. You can query testing data, monitor performance, and manage test automation within your AI workflow. The server is packaged as an MCP-enabled npm module and is started by installing the @smartbear/mcp package and providing credentials/tokens for the products you use. Tools are surfaced via prompts, prompts-driven inputs, and can be enabled or disabled by token presence. Typical use cases include retrievingBugSnag project data, opening Swagger portal information, triggering Pact-based contract tests, and coordinating Zephyr test cycles.
How to install
Prerequisites:
- Node.js 20+ and npm
- Access tokens for the SmartBear products you plan to use (BugSnag, Reflect, Swagger, QMetry, Zephyr, etc.)
Option A: Quick start (via npx)
- Ensure Node.js and npm are installed.
- Run the MCP server with npx (no local install required):
npx -y @smartbear/mcp@latest
- Follow prompts to provide tokens and configuration as needed; or configure .vscode/mcp.json as shown in the manual example to persist settings.
Option B: Local install (recommended for persistent usage)
- Install the package locally in your project:
npm install @smartbear/mcp --save-dev
- Create or edit your MCP configuration (see the sample in the repository) to include the smartbear server with the desired environment tokens.
- Start the server via your package manager/script, for example:
npm run mcp:start
Notes:
- If you are not using all SmartBear products, you can leave the corresponding environment variables blank to disable those tools.
Additional notes
Tips and common issues:
- Ensure Node.js 20+ is installed; MCP requires a compatible Node version.
- Tokens are optional; leaving a token blank disables the corresponding tool in MCP prompts.
- The MCP config is sensitive; avoid sharing your tokens. Consider using a secrets vault and referencing them in your env mapping as shown.
- When using npx, a network issue may require clearing npm cache or re-installing the package.
- If you update the MCP package, re-run the setup so that prompts reflect any new tools or configuration options.
- For VS Code Copilot setup, use the MCP: Add server command to scaffold the configuration and avoid manual editing where possible.
Related MCP Servers
mcp -azure-devops
An MCP server for Azure DevOps
BifrostMCP
VSCode Extension with an MCP server that exposes semantic tools like Find Usages and Rename to LLMs
better-notion
Markdown-first MCP server for Notion API - composite tools optimized for AI agents
thunderclient
Thunder Client MCP Server
xgmem
Global Memory MCP server, that manage all projects data.
PDFlow
Transform PDFs into structured data with AI-powered extraction + CLI + MCP + Web.
