# AutoSmoke

> AI-powered smoke testing platform. Runs a real browser through your critical user flows (signup, login, checkout, onboarding) after every deploy. Tests are written in plain English and self-heal when the UI changes — no selectors, no scripts, no maintenance.

## Product

AutoSmoke is a testing automation tool that uses AI to understand and interact with web applications the way humans do. Instead of relying on fragile element selectors that break when UI changes, AutoSmoke uses computer vision and contextual understanding to find and interact with elements. Tests are written in plain English, executed in real browsers, and self-heal when the UI changes.

## Audience

- Development teams building web applications
- QA engineers looking to reduce test maintenance burden
- Startups and teams without dedicated QA resources
- Teams using AI coding assistants who need fast, reliable testing
- Engineering leaders evaluating alternatives to Playwright, Cypress, or Selenium

## Capabilities

- AI-Powered Test Generation: Describe tests in plain English, AI generates robust test flows automatically
- Self-Healing Tests: Adapts to DOM changes without manual selector updates
- Visual Verification: Captures screenshots and visual diffs on every test run
- Scheduled Monitoring: Run tests on a schedule with notifications when something breaks
- Cross-Browser Support: Test across Chromium, Firefox, and WebKit
- CI/CD Integration: GitHub Action and GitLab CI templates for PR-based testing
- No-Code Test Authoring: Non-technical team members can write and maintain tests

## Limitations

- Designed for smoke and regression tests, not full QA coverage
- MFA / CAPTCHA / bot protection may require staging workarounds or self-hosted runner
- Currently web-only (no native mobile app testing)

## Primary URLs

- Home: https://autosmoke.dev/
- Features: https://autosmoke.dev/features
- Pricing: https://autosmoke.dev/pricing
- Contact: https://autosmoke.dev/contact

## Documentation

- [Getting Started](https://autosmoke.dev/docs/getting-started): Learn how to create your first test with AutoSmoke in minutes.
- [Writing Effective Tests](https://autosmoke.dev/docs/writing-tests): Best practices for writing reliable and maintainable tests.
- [Understanding Selectors](https://autosmoke.dev/docs/selectors): How AutoSmoke finds and interacts with elements on your page.
- [Handling Authentication](https://autosmoke.dev/docs/authentication): How to test pages that require login.
- [Setting Up Google OAuth Authentication](https://autosmoke.dev/docs/google-auth-setup): How to configure your Google OAuth credentials in AutoSmoke so your tests can run against authenticated pages.
- [CI/CD Integration](https://autosmoke.dev/docs/ci-cd-integration): Run AutoSmoke tests automatically on every pull request using GitHub Actions, GitLab CI, or any CI/CD platform.
- [Notifications & Alerts](https://autosmoke.dev/docs/notifications): Configure email, Slack, and webhook alerts to get notified when your tests fail, recover, or complete.
- [Troubleshooting](https://autosmoke.dev/docs/troubleshooting): Solutions to common issues when running AutoSmoke tests, including timeouts, authentication problems, element detection, and CI failures.

## Blog

- [After TanStack: A Practical Defense Against the Next npm Supply Chain Attack](https://autosmoke.dev/blog/surviving-npm-supply-chain-attacks-after-tanstack): The TanStack compromise was the latest in a long run of 2025–2026 npm supply chain attacks. The pattern is consistent, and so is the playbook for surviving the next one. Here's what actually works.
- [Beyond Playwright: Why Agentic AI Is Eating End-to-End Testing in 2026](https://autosmoke.dev/blog/agentic-ai-testing-beyond-playwright-2026): Playwright just overtook Selenium for the first time — and a new wave of agentic AI testing tools is already making script-based frameworks feel like the next legacy. Here's what changed in 2026, and why post-deploy smoke testing will never look the same.
- [The Vercel Breach: When Your AI Assistant Is the Supply Chain](https://autosmoke.dev/blog/vercel-april-2026-security-breach): On April 20, 2026, Vercel disclosed that attackers accessed internal systems through a compromised AI tool's OAuth connection to Google Workspace. The blast radius started with a Roblox cheat script. Here's the chain—and what it means for anyone wiring AI tools into their SaaS stack.
- [Vercel's Dubai Datacenter Failure: When a Single Region Takes Down Global Builds](https://autosmoke.dev/blog/vercel-dubai-datacenter-outage-march-2026): On March 2, 2026, a datacenter failure in Vercel's Dubai (dxb1) region didn't just break deployments in the Middle East—it caused build failures worldwide for any project using Middleware. Here's what happened and what it means for your deployment strategy.
- [When the Lights Go Out: Major Software Outages of 2024–2026](https://autosmoke.dev/blog/major-software-outages-2024-2025): From CrowdStrike's 8.5 million blue screens to AWS losing DNS for 15 hours—a breakdown of the biggest software outages in recent memory and what they reveal about modern system fragility.
- [The Critical Role of Testing and Monitoring in the Age of AI Development](https://autosmoke.dev/blog/testing-monitoring-ai-development): As AI accelerates software development, the importance of robust testing and monitoring has never been greater. Here's why quality assurance is more critical than ever.

## Full Content

- All documentation as a single file: [llms-full.txt](https://autosmoke.dev/llms-full.txt)
- All blog posts as a single file: [llms-blog.txt](https://autosmoke.dev/llms-blog.txt)
- RSS feed for blog: [feed.xml](https://autosmoke.dev/blog/feed.xml)

## Comparison Pages

- [AutoSmoke vs Checkly](https://autosmoke.dev/compare/autosmoke-vs-checkly)
- [AutoSmoke vs QA Wolf](https://autosmoke.dev/compare/autosmoke-vs-qa-wolf)
- [AutoSmoke vs Playwright](https://autosmoke.dev/compare/autosmoke-vs-playwright)

## Contact

- Email: contact@autosmoke.dev
- Website: https://autosmoke.dev