Skip to main content
Glama
Browser Core

Browser Core

by ShuhLab

Browser Core — browser automation with human control

Connect AI agents to a browser you own. Choose your local Chrome or a personal Linux server.

Browser Core is an MCP-connected browser automation service by Shuh Lab. It opens pages, reads accessible controls, clicks, fills forms, extracts information, takes screenshots and saves encrypted browser profiles. When a login requires a password, CAPTCHA, 2FA or OAuth approval, the agent pauses for a user to briefly take control of that same session, then resumes.

Status: free, invitation-only beta. Not a public paid cloud product.

Current capabilities

Feature

Local Mac

Personal Linux BYO

Hosted commercial cloud

MCP browser automation

Tested

Tested

Not available

Encrypted, reusable browser profile

Yes

Yes

Not available

Manual login / CAPTCHA / 2FA

Native Chrome or controlled window

Short, temporary remote live view

Not available

Connect to existing, individually approved Chrome tabs

Yes

No

Not available

Infrastructure provided by the user

Local computer

Your own server

Not available

Manual remote viewing is ONLY for short authentication or verification steps. It is not intended for normal browsing; use your own computer when you need comfortable interactive work.

Related MCP server: chrome-mcp

Getting started

  1. Request an invitation via https://t.me/OpsTetherBot?start=browsercore .

  2. Connect the MCP endpoint using OAuth in a compatible client.

  3. Choose Local or BYO and pair a worker with a one-use code.

  4. On Chrome, grant the specific tab you want the agent to operate.

  5. Enter sensitive login details only in the real browser, never in chat.

The current Mac installer is unsigned, not notarized, and requires Python 3.10–3.13. Chrome extension installation may require Developer Mode. Windows, Apple Silicon and large-scale commercial hosted Cloud are not yet release-certified.

Why consider Browser Core?

  • Work in your existing, explicitly approved Chrome tab instead of copying a profile into an agent cloud.

  • Choose a local computer or your own Linux host; no mandatory commercial browser-minute subscription for these two modes during the beta.

  • Human and agent can cooperate in the same remote context during an authentication challenge.

  • Persistent sessions and encrypted browser profiles, with explicit permission boundaries.

These are our current product design choices, not claims of industry-first features. Manual takeover exists in Browserbase and other platforms. TinyFish offers a mature managed Chromium, CDP and proxy service that this beta does NOT replace at scale.

See detailed, sourced product positioning in docs/alternatives.md.

Documentation

Distribution, access, and payment

Invite-only beta access is free. Payment activation, online checkout and shared hosted Cloud capacity are disabled. This repo contains documentation and release metadata only; the service backend is proprietary, not open source.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    Related MCP Connectors

    Related MCP Servers

    • A
      license
      Not graded
      quality
      B
      maintenance
      Enables any coding agent to use the user's existing Chrome profiles and logins to open tabs, read pages, fill forms, click through flows, and take screenshots.
      81 npm
      MIT
    • A
      license
      Not graded
      quality
      A
      maintenance
      Enables AI assistants to drive the user's real, already-logged-in Chrome browser locally, so they can read and navigate pages, click, type, fill forms, take screenshots, and inspect cookies, tabs, console, and network traffic without headless browsers or remote-debugging ports.
      1 npm
      ISC