Kobiton Claude Code plugin writes end-to-end tests, uploads builds, reserves real devices, runs tests, and analyzes results from your IDE. The agent uses your project as context to write tests from existing requirements. Type test prompts in plain English and the agent generates code in your chosen language using open source Appium that can be reviewed and customized. Build upload and test execution occur without leaving the IDE. Tests execute in real-time on physical devices with instant feedback. Test results include agent-generated explanations that read logs, view screenshots, confirm assertions, and detail what passed or failed.
Kobiton converts manual test sessions captured on real devices into production-ready Appium scripts without scripting. The AI reads taps, swipes, and inputs from manual sessions and exports complete Appium packages in Java, Python, NodeJS, or C#. Generated scripts include full source code with boilerplate, device configuration, and element locators pre-configured. Scripts require only API key insertion to run. Execute test suites directly from IDE with right-click run command. Scripts run on physical devices with live tracking in Sessions list. Manual test to executable Appium script conversion eliminates element hunting and supports Android and iOS device labs.
Test biometrics, image injection, location mocking, and device metrics on physical iPhones. Simulate fingerprint and Face ID authentication to test authentication flows without physical interaction. Camera button intercepts enable image injection to test receipt scanning, ID verification, boarding passes, and barcode functionality. Set device GPS location anywhere globally to test geofencing, store locators, delivery zones, and region-locked features without physical travel. Real-time metrics measure memory, CPU, network, battery, and temperature on physical devices during test execution.
Run one manual test and replay it across multiple devices without scripting or maintenance. Recording captures every tap and input automatically in the background during application navigation. Recorded sessions convert to automated test steps with screenshots that execute across selected Android or iOS devices simultaneously in parallel. One manual session generates automated tests for multiple devices without automation engineering.
Session Explorer automatically captures and surfaces issues from test sessions. Screen load times are measured with threshold violations flagged. Screens exceeding 2 second threshold are marked red and clickable on timeline. Permanent links to exact test steps can be copied and shared in Slack or Jira for developer reproduction with full context including logs, payloads, and metrics. Built-in element inspector surfaces locators directly from sessions without separate Appium inspector installation. CPU, memory, network, battery, and temperature metrics correlate to exact timeline moments. One-click bug filing to Jira includes pre-populated session link, OS, resolution, device info, logs, crash logs, and Appium logs. Network captures export as HAR files for root cause analysis, security review, and compliance auditing.