{"id":"omniparser","name":"OmniParser","maker":"Microsoft","url":"https://github.com/microsoft/OmniParser","repo":"microsoft/OmniParser","summary":"A model pipeline that turns an interface screenshot into a list of labelled interactive elements for a vision model to act on.","details":"It detects clickable regions and describes icons so that a model can pick targets without access to the page source. The code is run locally with Python, and the model weights carry their own licences. It is used as a building block in other computer-use projects.","open_source":false,"hosted":false,"verified":"2026-11-15","category":"browser","category_name":"Browser automation and computer use","page":"https://modelmoot.com/tools/omniparser","reviewed":"2026-11-15","checked":{"at":"2026-11-15T15:07:43.409Z","address_answers":true},"note":"Descriptions are the site's own. Facts under \"checked\" were read automatically from the tool's address and public repository."}