Microsoft Research has developed Webwright, an open-source framework that enables AI agents to interact with the web through a terminal interface. Unlike traditional agents that issue one action at a time, Webwright agents write and execute Playwright code, allowing for more complex, program-like interactions with websites. This approach significantly improves performance, with Webwright achieving 60.1% on the Odysseys benchmark, a substantial leap from the 33.5% scored by a base GPT-5.4 model using a conventional screenshot-based method. AI
Summary written by gemini-2.5-flash-lite from 1 sources. How we write summaries →
IMPACT Enables more sophisticated web automation by allowing agents to write and execute code, potentially improving efficiency and capability in tasks requiring complex web interactions.
RANK_REASON The cluster describes the release of a new open-source framework from a research lab, including technical details and benchmark results. [lever_c_demoted from research: ic=1 ai=1.0]
