{"id":"browserstack-mcp-server","name":"BrowserStack","af_score":62.2,"security_score":67.2,"reliability_score":53.8,"what_it_does":"BrowserStack's official MCP server that enables agents to programmatically manage and execute test cases, run manual/automated tests on real devices and browsers, debug test failures, and perform accessibility testing through natural language prompts integrated into IDEs like VSCode, Cursor, and Claude Desktop.","best_when":"Teams need to integrate comprehensive cloud-based testing capabilities into their development workflow, want AI-assisted debugging of test failures, or need accessibility compliance without context switching between tools.","avoid_when":"Your organization has strict data residency requirements, needs zero-latency local testing, or lacks a paid BrowserStack account (free tier limited to open-source projects).","last_evaluated":"2026-04-04T22:16:01.016737+00:00","has_mcp":true,"has_api":false,"auth_methods":["API credentials (username + access_key)"],"has_free_tier":true,"known_gotchas":["Credentials must be stored in environment variables, not hardcoded in prompts","No explicit error recovery or retry guidance in documentation","Tools require valid BrowserStack session IDs and project identifiers; agent must handle state management","No documented idempotency; duplicate operations may create duplicate test results or sessions","MCP server launch requires Node.js >=18.0; version mismatch could cause silent failures","Rate limits not explicitly documented; agent may exceed limits without clear feedback"],"error_quality":0.0}