Auto WebMCP

Auto WebMCP turns forms, standalone inputs, supported widgets, and searches into structured WebMCP tools for compatible browser agents. An agent can discover available tools and fill several fields in one call. Supported controls include text fields, text areas, select menus, multi-selects, radio groups, checkboxes, dates, times, numbers, and ranges. Supported input groups work outside HTML form tags, and tools update when pages change. Widget tools open supported disclosures and select currently available options. Discovery does not automatically open controls. Search tools support eligible same-origin GET search forms and Willhaben marketplace searches with supported area, price, condition, and sorting filters. Support depends on each website. Fill tools do not call form submission. Widget actions can trigger website behavior. Search tools navigate to a results URL, sending search values to the destination website. Existing page-provided WebMCP tools are preserved. Hidden, disabled, read-only, and file-upload controls are excluded from filling. The extension runs locally in Chrome with no developer-operated backend, analytics, advertising, or telemetry. Your chosen agent and the websites you use have their own data practices. Source code and privacy policy: https://github.com/pauloportella/auto-webmcp-chrome Requires Chrome 149 or later and a compatible WebMCP browser agent or client.