Pagerunner, a free MCP server from Enreign, automates Chrome browser profiles for various tasks. Built with Rust, it connects to existing user data like cookies and saved passwords. This tool is distributed via Homebrew, Cargo, and GitHub binaries, making it accessible for developers. It enables AI clients to interact with real browser environments, supporting a range of automation needs from web testing to data gathering.

What does Pagerunner connect to?

Pagerunner connects directly to real Chrome browser profiles. It uses existing user data, including cookies, saved passwords, and browsing history, to operate within authenticated states. This allows it to perform actions as if a human user were present. The program functions as an MCP server, specifically designed for browser automation. It requires users to close any Chrome window using a profile before initiating a Pagerunner session on it, due to Chrome's profile directory locking.

Core features and tools

Pagerunner exposes a set of callable tools for browser automation. It offers 27 tools for a complete toolkit, including navigate, click, type, fill, screenshot, and evaluate. Specific tools used in workflows are screenshot(), navigate(), get_content(), and click(). These tools cover sessions, navigation, interactions, snapshots, checkpoints, KV store, site intelligence, secrets, and recordings, enabling diverse automation tasks within Chrome.

Who is it for?

Pagerunner is designed for AI agents and teams needing to automate browser interactions. It supports frontend development, competitor analysis, and quality assurance workflows. Teams can use it for visual regression testing or automating acceptance criteria walkthroughs. AI agents benefit from its ability to read project management tickets within persistent browser profiles, gathering context for coding. Skip this if you need to automate browsers without using existing Chrome profiles.

How to get started

The authentication model relies on using existing Chrome profiles, which contain authenticated states. Pagerunner encrypts this saved state using AES-256-GCM, with the key stored in the macOS Keychain or set via the PAGERUNNER_DB_KEY environment variable on Linux. It supports the stdin/stdout transport protocol for exposing its MCP server.

Use cases

• Automating frontend development tasks, such as editing CSS, navigating, taking screenshots, and comparing against design specifications.
• Monitoring competitor offers, feature pages, and changelogs on a schedule to compile morning briefings.
• Performing visual regression testing by screenshotting feature branches at various breakpoints, pixel-diffing against baselines, and flagging regressions before merges.

Pagerunner provides an effective solution for automating Chrome browsers, connecting AI agents with real user profiles and their data. Its Rust foundation ensures performance, while encryption secures sensitive profile information. It is compatible with various AI clients, making it a versatile tool for development, testing, and monitoring.