March 2026 Summaries

Cypress has introduced cy.prompt, an AI-powered feature that allows users to write test automation scripts in plain English, eliminating the need for manual coding. Initially launched as an experimental feature at CypressConf 2025, cy.prompt is now in beta and fully integrated into Cypress 15.13.0, removing the need for prior experimental configuration. The tool offers enhanced features such as positional context for more precise element targeting, visible generated code for easier debugging without sourcemaps, and expanded assertions for more comprehensive testing. It also supports interactions like keyboard navigation and hover states, automatic scrolling for elements out of view, and self-healing selectors that adapt to changes in the DOM. Designed to bridge the gap between test conception and execution, cy.prompt aims to improve testing coverage and reliability, with ongoing efforts to determine future pricing based on user feedback during the beta phase.

Cloud MCP, a new beta feature from Cypress, is a remote Model Context Protocol server designed to enhance AI coding assistants by connecting them directly to Cypress Cloud, thus addressing the context gap in AI-driven development. This feature allows AI assistants to access real-time test run data, including run statuses, failure details, error messages, and stack traces, thereby eliminating the need for manual debugging processes. By integrating seamlessly with AI tools like GitHub Copilot and OpenAI’s Codex CLI, Cloud MCP enables developers to perform root-cause analysis and investigate flaky tests efficiently without context switching. The server is accessible for free on all Cypress Cloud plans, ensuring that AI-assisted debugging is available to all developers using Cypress. As the platform evolves, further enhancements are being explored to provide AI agents with even more comprehensive data, aiming to transform AI assistants into fully autonomous development agents.

Cypress has enhanced its cy.prompt feature to support waiting for network requests, allowing testers to reference defined aliases directly within prompt instructions, thereby streamlining the testing process without exiting the flow. This improvement eliminates the previous need to split test intent across multiple lines when coordinating network activities, thus maintaining readability and coherence in AI-assisted test writing. By defining the intercept and alias with cy.intercept and .as() outside of the cy.prompt block and referencing the alias using the @ prefix, testers can ensure seamless execution. While this update currently handles basic network coordination by waiting for a named intercept, it resolves a significant friction point without altering existing intercept patterns. Cypress encourages feedback to further develop cy.prompt's capabilities, particularly for more complex asynchronous sequences.

Studio AI is an enhancement to Cypress Studio that aims to improve release confidence by facilitating stronger validation with less friction in UI testing. It works by transforming visible changes in the UI into assertions using AI-generated Smart Recommendations, which are based on observed differences in the DOM after each recorded interaction. These recommendations are integrated into the testing workflow, allowing users to accept, reject, or modify them as needed, with the option to view DOM snapshots to understand the changes being validated. Studio AI prioritizes stable selectors and offers configuration options for custom conventions, while safeguarding sensitive information by excluding certain data types from analysis. It requires Cypress version 15.11.0 or later and a Cypress Cloud account, with Smart Recommendations available at no additional cost during the beta phase. By connecting to Cypress Cloud, Studio AI supports broader testing workflows, encouraging consistent validation practices that enhance the depth of test coverage and predictability of release confidence.