September 2026 Summaries

Sauce Labs has introduced its Hosted Model Context Protocol (MCP) server, which connects AI assistants such as Claude, Cursor, Windsurf, and other compatible clients directly to its testing platform through a hosted HTTP service. The platform enables agents to generate and export test cases, find and control real devices, install app builds, run and schedule tests, collect screenshots and logs, inspect artifacts, execute shell commands, and configure or capture network activity using natural-language prompts. Sauce Labs positions the offering as distinctive because it combines AI test authoring and real-device execution in one maintained service, avoiding locally installed servers and manual dashboard workflows. The company describes use cases for developers diagnosing crashes, SDETs creating tests from Jira tickets, QA leads validating releases across device configurations, and support teams reproducing customer issues, with setup requiring an MCP client to connect to mcp.saucelabs.com using existing Sauce Labs credentials.

Sauce Labs argues that many customers use only its virtual device testing capabilities while overlooking a broader release-assurance platform that includes real-device testing, AI-assisted test authoring and analytics, visual regression, accessibility testing, app distribution, and production error reporting. The company positions its AURA platform as a unified Requirements-to-Monitor workflow in which AI agents assist with defined tasks under human review, and it claims that consolidating these functions can reduce tool switching, duplicate spending on products such as Applitools, BrowserStack, Crashlytics, or Sentry, and improve feedback loops between development, testing, beta releases, and production incidents. Sauce Labs cites customer and third-party-reported gains in engineering capacity, release speed, issue resolution, test coverage, crash reduction, and return on investment, while emphasizing its device cloud of more than 10,000 real devices and support for thousands of OS combinations. It also highlights compliance certifications and integrations, presenting the platform as a single data layer intended to connect software requirements, test results, visual and accessibility findings, and production failures.