RunBrowser is a Chrome extension plus a Node.js CLI/MCP tool. Instead of launching a fresh headless or standalone Chrome instance like Playwright MCP, it connects to the real Chrome tabs the user is already using. This allows agents to work directly with existing login sessions, cookies, and extensions, making it well suited to web tasks that require account context, human intervention for CAPTCHAs, or sensitivity to anti-bot detection.
Its architecture uses the extension to connect to tabs via chrome.debugger and open a WebSocket relay on local localhost:19988; the CLI or MCP then sends CDP commands through that relay. Functionally, it covers 50+ commands including navigation, clicking, typing, waiting, screenshots, accessibility snapshots, tab and iframe management, eval, and raw CDP. Accessibility snapshots output an element tree with @ref markers, which is more suitable than screenshots for low-cost LLM parsing. Site commands are TypeScript plugins that can wrap site-specific scraping workflows into structured JSON output; community commands can be installed from GitHub repositories.
The main text does not disclose pricing, plans, account systems, or paid limitations. The page mentions GitHub stars, a community command repository, and PR contributions, but does not clearly state the main project license, so its full open-source status cannot be directly determined. By default, it runs locally with no remote server, no account, and no telemetry; it also supports exposing the relay for remote access via a host and token.
The main advantage is that it can reuse a real browser environment, including login state and extensions. Its CLI, MCP, and TypeScript command mechanism are friendly for agent developers. It also has a clear human-in-the-loop workflow, allowing users to handle CAPTCHAs, cookie pop-ups, and unexpected pages in real time. The downsides are that first-time setup requires installing the CLI, loading the extension, and authorizing the target tab. The main text currently only explicitly mentions Chrome. Information on commercial support, SLAs, version compatibility, and pricing is also insufficient. Remote access is possible, but users must take responsibility for network exposure and token management.
RunBrowser is suitable for agent engineers, automation script developers, web data extraction use cases, and teams that need to operate with logged-in browser sessions. It is less suitable for non-technical users who do not want to configure a local environment. Access from China is not discussed in the main text; given that npm, GitHub community commands, and the MCP client ecosystem may be affected by network conditions, real-world usability would need to be tested. Alternatives include Playwright MCP, BrowserUse, and Agent Browser.
β This review is compiled from public sources and does not constitute a purchase recommendation. Verify all facts on the vendor's official site. Verify on runbrowser.com official site.
runbrowser.com is an United States Dev Tools provider. TG4G tracks its product information, an overall rating of 8.0/10, and a China-accessibility score of China direct-connect friendly. Click "Visit Official Site" to reach runbrowser.com directly.