Optics

•
72
•

Automate and execute self-healing end-to-end tests across mobile, web, smart TV, and Bluetooth devices using multi-strategy locators and AI agent integration.

Optics Framework is an open-source test automation engine designed to reliably drive mobile, web, smart TV, and BLE devices without brittle failures. By employing a fallback locator ladder that cascades through native XPaths, visible text, multi-engine OCR, visual template matching, and LLM-assisted self-healing, it ensures tests adapt seamlessly when interface elements change. Test suites can be authored as declarative CSV/YAML files, captured through an interactive terminal recorder, or embedded into code via Python and Robot Framework. Optics also functions as an MCP server, allowing AI agents to inspect live device states and autonomously execute actions.

Key Features

01Cross-platform support covering Android, iOS, web browsers, Android TV, Tizen, webOS, and Bluetooth HID devices.
02Self-healing locator ladder falling back through native XPath, text, OCR, image template matching, and AI recovery.
0365 GitHub stars
04Data-driven test authoring using plain CSV or YAML files alongside Python SDK and Robot Framework integrations.
05Built-in Model Context Protocol (MCP) server exposing keywords as actionable tools and device state as inspectable resources.
06Interactive terminal recording mode for live exploratory testing and automatic test case generation.

Use Cases

01Testing locked-down production devices non-intrusively using Bluetooth HID mouse and keyboard emulation paired with vision detection.
02Executing cross-platform end-to-end regression suites across web, mobile, and smart TV platforms with a unified syntax.
03Empowering AI agents to interact with and test real hardware devices or mobile apps via standard MCP tooling.