Write what the test should do in plain English. The Text2Test Native model, trained specifically on QA context, interprets your words, compiles them into deterministic steps, and runs them consistently across any platform.
No syntax. No formatting. No learning curve.
Write your instruction in plain text. Paste a Slack message, a Jira ticket, or a one-line spec. T2T Autopilot hands it to an LLM that reads the screen, reasons about the flow, and executes dynamically. Best for complex branching flows where rigid steps fall short.
Paste an existing Playwright script directly into the sequence. Text2Test runs it exactly as written, alongside every other step in the test case. Don't have one yet? Generate a context-aware Playwright script from a test case, then inject it.
Upload a reference screenshot at any point in your sequence. Describe what should match in plain text. Text2Test compares the current screen against your reference and tells you whether it does. No pixel thresholds. No config files.
Call any API endpoint inline. Assert on status, body, headers, or response time. Save the request to your workspace and reuse it as a data source in plans, a step in another test case, or a standalone API call. One request, three jobs.
Each test case is a sequence of steps. Each step can be a different type, combined in any order. A login can run in plain text. The next step can hit an API to seed test data. The third can verify a screenshot. The fourth can inject a script for a tricky form. The last can hand control to Autopilot for the final flow.
Same sequence. Same run. Same report.
