The ChatGPT for end-to-end browser testing.
We are opening up Bug0 Studio v0.1 in research preview. This is the internal tool our FDE team uses to turn natural language and video into clean, reliable Playwright tests.
Demo from the founder
What it does
1. Understands visual context
Studio processes video recordings of real user flows. You can record your browser tab, upload an mp4/webm, or type a natural language description. The model sees UI state, user intent, and dynamic elements that text-only LLMs usually miss.
2. Validates logic before code
After analyzing the video, Studio extracts ordered steps. You can edit, add, or remove steps. This avoids black-box output and keeps full control over the logic.
3. Runs tests in a live cloud browser
Studio spins up a live execution environment. Left side shows AI reasoning. Right side shows the test running in a real browser. Scripts are aligned to actual app behavior, not static HTML.
4. Generates robust Playwright scripts
Studio outputs clean, intent-based code using resilient selectors like getByRole. No brittle nth-child paths. No vendor lock-in. All tests run in your own CI.
5. Handles authentication cleanly
Paste your Playwright storageState.json to skip login steps and test deep-link flows instantly. Base URLs and credentials stay in your browser’s localStorage. Nothing stored on our servers.
What’s inside v0.1
-
Video-first test generation
-
Natural-language to Playwright
-
Step-level validation
-
Live cloud browser execution
-
Robust selectors
-
Storage state support
-
Standard Playwright output
What’s coming next
We are exploring features like smarter branching flows, deeper cloud browser controls, and tighter CI integrations. More updates soon.
Studio runs on Passmark, our open-source testing engine. It handles discovery, self-healing, and deterministic Playwright execution. Read why we open sourced it.
Try it
Public preview is live.
Report issues or feature requests in Discord: go.bug0.com/discord.
