browse.camCapture beta

01 / Capture

Your browser.
Your signed-in page.

No browse.cam or dots login needed. Open any website you can access, then choose a tool below.

Nothing captured yet.

Selected page

Choose a tab

Click the browse.cam icon on a website.

How to teach a course

Mark values and controls, draw arrows, or trace a mouse path. Switch to Use page for your own live clicks. Finish to name variables and save a local course draft.

One page per draft: save before navigating or reloading. No screenshots, model calls or typed input values in the teaching loop. Drafts do not execute actions. The popup closes when teaching starts so you can use the on-page toolbar.

Teach this cam a source

Select a visible value or button to save its selector, role and position. The selection layer receives your click; it does not invoke the selected control. Escape cancels.

Select the source, then reopen this popup. Embedded frames and shadow-root internals are not supported.

Live reading samples one selected element at most once per second and skips unchanged values. It stops on tab, URL or document changes. Keep the popup or side panel open. Video and canvas need pixel capture; this mode does not read their frames. Nothing uploads automatically.

    
Leave blank for the viewport and page text. A focus region must fit entirely on screen; only its screenshot and text are saved.

    
Measure capture performance
Compare the viewport with your focus region: one warmup and five samples each. Keep the page still for about 10 seconds. The report includes timings and sizes, without page images or text.
Measures capture and image processing only. OCR and page loading are separate tests in the web benchmark suite.

    
Replay an existing course
Import its course JSON to collect the declared DOM selectors. The tab must match an exact course URL.
No course selected · capture page text only