No user-submitted related servers found.
Scored across 23 tools
Most tools have clearly distinct purposes (type, press, select, wait_for, cookies, dialog, network), and descriptions explicitly resolve the trickier pairs — 'visual' is framed as the cheap orientation screenshot vs 'screenshot' for annotated/full-page captures, and 'snapshot' (a11y refs) vs 'dom' (raw DOM queries). The remaining ambiguity is mild: cursor/press/drag all perform interaction, and visual/screenshot/snapshot all 'look' at the page, but the guidance on when to use each is clear enough that an agent is unlikely to misselect.
Everything is lowercase snake_case, but the lexical convention is mixed: some tools are verbs (scroll, navigate, type, press, evaluate, fill_form, upload_file, wait_for) and others are bare nouns (visual, cursor, status, tabs, screenshot, dom, console, network, cookies, dialog, debugger). Several tools are also multiplexed behind an 'action' parameter rather than having action-specific names. Readable, but no predictable verb_noun pattern.
At 23 tools it sits at the heavy end for a browser automation server, and a few entries feel partially redundant (visual vs screenshot, snapshot vs dom). Each still has a real role in the workflow, so it is not bloated, but the surface is broader than necessary for most agent tasks.
The browser lifecycle is well covered: connection status, tab management, navigation, interaction, form filling, file upload, async waiting, plus diagnostics (console, network, cookies, dialogs) and power features (evaluate, emulate, debugger). Gaps are minor and workaroundable — no dedicated hover tool, no direct localStorage/sessionStorage access, and no download or permission handling (evaluate can cover most of these).