{"id":"webvoyager","name":"WebVoyager","maker":"Zhejiang University and Tencent AI Lab researchers","url":"https://github.com/MinorJerry/WebVoyager","repo":"MinorJerry/WebVoyager","summary":"A web agent and benchmark of tasks on live websites, judged with a multimodal model.","details":"The benchmark contains tasks on a few dozen real sites, and the repository includes a screenshot-based agent and an automatic evaluator that asks a vision-language model to judge success. The code is open source under Apache 2.0.","open_source":true,"hosted":false,"also":["browser"],"verified":"2026-11-15","category":"benchmarks","category_name":"Agent benchmarks","page":"https://modelmoot.com/tools/webvoyager","reviewed":"2026-11-15","checked":{"at":"2026-11-15T17:14:07.328Z","address_answers":true},"note":"Descriptions are the site's own. Facts under \"checked\" were read automatically from the tool's address and public repository."}