What is this?
Expect is a testing tool that runs your AI agent's code against a real browser, so you can verify your agent behaves correctly in an actual browser environment instead of relying on mocks.
Expect is a testing tool that runs your AI agent's code against a real browser, so you can verify your agent behaves correctly in an actual browser environment instead of relying on mocks.Is this a CLI tool or a library I import into my code?It ships as a CLI, living in the apps/cli package of the repo's Turborepo monorepo, rather than being a library you import directly.
It ships as a CLI, living in the apps/cli package of the repo's Turborepo monorepo, rather than being a library you import directly.Does Expect test more than just browser-based agent behavior?Based on the description, Expect's focus is specifically on verifying agent code against a real browser, so non-browser agent paths don't appear to be its target use case.
Based on the description, Expect's focus is specifically on verifying agent code against a real browser, so non-browser agent paths don't appear to be its target use case.Who is this built for?It's aimed at developers building AI agents that act inside web browsers and who need automated, real-browser end-to-end checks of that behavior.
It's aimed at developers building AI agents that act inside web browsers and who need automated, real-browser end-to-end checks of that behavior.Why test in a real browser instead of just mocking it?Agent behavior like DOM interactions, rendering, and timing can only be reliably verified against an actual browser instance, not a simplified mock.
Agent behavior like DOM interactions, rendering, and timing can only be reliably verified against an actual browser instance, not a simplified mock.What does the name 'Expect' hint at?The name suggests assertion-style 'expect' checks, similar to common testing libraries, used to verify an agent's actions in the browser.
The name suggests assertion-style 'expect' checks, similar to common testing libraries, used to verify an agent's actions in the browser.Is the project built as a single package or a monorepo?It's a Turborepo-based monorepo with a dedicated CLI app and shared build, lint, and test pipelines managed through pnpm and Turbo scripts.
It's a Turborepo-based monorepo with a dedicated CLI app and shared build, lint, and test pipelines managed through pnpm and Turbo scripts.Does the project have CI in place?Yes, the repo runs code-quality, end-to-end, and release workflows via GitHub Actions, including a dedicated e2e.yml pipeline.
Yes, the repo runs code-quality, end-to-end, and release workflows via GitHub Actions, including a dedicated e2e.yml pipeline.What are similar projects?playwright, testsprite-cli, e2e, SeleniumBase, assay, midscene, open-agents, agentic-playwright
playwright, testsprite-cli, e2e, SeleniumBase, assay, midscene, open-agents, agentic-playwrightWhich topics is it in?turbo, end-to-end-testing, test, e2e-testing, test-automation, pnpm, monorepo, browser-automation, ci, testing, browser, agents
turbo, end-to-end-testing, test, e2e-testing, test-automation, pnpm, monorepo, browser-automation, ci, testing, browser, agents