agent-in-browser
Related Servers
Alternatives to agent-in-browser
No user-submitted related servers found.
Related Servers
- AlicenseNot gradedqualityCmaintenanceEnables AI agents to control a real browser via MCP tools for web navigation, interaction, page inspection, tabs, forms, files, sessions, and browser automation.MIT
- AlicenseNot gradedqualityAmaintenanceEnables AI agents and human operators to launch and drive isolated, persistent browser profiles — each with its own cookies, fingerprint, and proxy — from any MCP client. Exposes tools for navigating, clicking, typing, extracting content, taking screenshots, evaluating JavaScript, managing tabs and cookies, and waiting on page events.MIT
- AlicenseNot gradedqualityAmaintenanceEnables AI agents to drive isolated local browser profiles, each with its own persistent cookies, login state, fingerprint, and proxy, so authenticated workflows carry across sessions. Exposes tools for navigation, clicking, typing, extraction, screenshots, JS evaluation, tab and cookie management, and raw CDP commands to any MCP client.MIT
- AlicenseBqualityCmaintenanceEnables AI agents to perform low-latency web automation with a persistent headed Chrome/Chromium instance, including navigation, DOM interaction, content extraction, screenshots, tab management, and batch/code execution via 18 MCP tools.20MIT
- AlicenseNot gradedqualityBmaintenanceEnables AI to control a real browser through MCP tools for clicking, typing, navigation, screenshots, and more. It supports a follow mode that tracks the active tab, plus fixed mode for controlling specific tabs.18 npmMIT
- AlicenseNot gradedqualityAmaintenanceEnables MCP-compatible AI agents to navigate, interact with, and automate the web through an anti-detection browser toolset, supporting multi-tab workflows, session persistence, downloads, extraction, and token-efficient accessibility snapshots.899 npmMIT
TDQS
Scored across 23 tools
Tools split cleanly along interaction vs. inspection lines, and descriptions actively cross-reference each other (get_text vs snapshot vs screenshot, click vs hover, type vs fill_form vs press_key, browser_status vs tabs_list vs browsers_installed). The only soft spot is the read-inspection cluster (snapshot, find, get_text, evaluate), which could occasionally be confused, but each is scoped to a different need.
Everything is snake_case, which is good, but the ordering pattern mixes verb-first (get_text, fill_form, press_key, wait_for) with noun-first (browser_close, tabs_list, tab_new, console_messages, network_requests). Prefixes also drift between singular `tab_*` and plural `tabs_*`, and `browser_*` vs `browsers_*`, making the scheme only loosely predictable.
23 tools is heavy for the stated scope and pushes into the crowded band; a browser agent genuinely needs much of this surface, but some tools (browsers_installed, resize, scroll, find) are arguably niche relative to core navigate/snapshot/click/type. It is defensible but not tightly scoped.
Coverage is broad: navigation and history, tab lifecycle, trusted input, form filling, keyboard, scrolling, screenshots, viewport emulation, console/network diagnostics, JS evaluation and waiting. Gaps exist around file upload/download, dialog handling, cross-origin iframe targeting and full-page capture, but most are workable via evaluate or press_key.