About the service
Agentic automation infrastructure and orchestrator from the German AskUI team: the agent looks at the screen and works the mouse and keyboard the way a person does, so it needs no selectors and no access to an application's internals. A task is written in plain words in a Markdown file kept in your own repository — the vendor stresses that what gets stored is the intent rather than a path through one particular layout, which is why the same file survives a redesign or a port to another platform. It comes in three parts: AskUI Desktop, where tasks are written, watched and replayed; the AskUI CLI, the same engine without a window, for pipelines, schedules or a box in the corner; and the AgentOS runtime, which hands the agent a keyboard, mouse and screen on Windows, macOS, Linux, Android, in browsers and on bench hardware. Every run leaves a verdict, a screenshot per step and the full transcript. You can bring your own model and keep it on your own hardware — the vendor explicitly notes that screenshots never leave your network, claims ISO 27001 and GDPR, and offers an appliance and air-gapped deployment. It is used above all for validating and documenting interfaces, document work and vendor portals. AskUI publishes no price list: quotes are written per organisation around the number of concurrent agents and the deployment model, with no per-seat and no per-task fees, so the exact terms are a question for the seller.