Browser Mesh
Drives a real Chrome browser through MCP, supporting both an everyday Chrome profile via an extension and Native Messaging and a dedicated profile launched by the daemon. Exposes tools for tabs, navigation, accessibility snapshots, screenshots, clicking, typing, form filling, keyboard input, scrolling, hovering, JavaScript evaluation, dialogs, console logs, browser lifecycle, and window management.
Enables cross-device control of a host's browser over a Tailscale network, allowing another trusted device on the tailnet to act as a controller.
This server does not currently have a deployable release on Glama. Follow the repository's setup instructions to install and configure it for your MCP client, including any required credentials and local dependencies.
Browser Mesh
Computer use for the browser: AI agents work in a real Chrome — the one you are signed in to, or a separate one — with every site and permission under your control.
Browser-level computer use. Agents do not just read and click page content: they also handle Chrome's own interface — the "Save password?" bubble, site permission prompts, downloads, HTTP sign-in dialogs and the account chooser — in your everyday, signed-in Chrome. Because it is only the browser, the control is finer than general computer use: grants per site and per scope, a person confirms anything that raises privileges, one-click emergency stop, a local audit log, and saved passwords that agents can use but never read.
A browser mesh across your devices. An agent on any device in your Tailscale network can drive the browser on another. Each device is identified by
tailscale whois, grants are per device, and the host's owner approves a device the first time it connects.Remote help, shown step by step. Help someone who installs only the extension: an end-to-end encrypted relay opens on your computer only for the session, they watch everything, confirm the risky steps themselves, and the session has a time limit.
Status: early (0.1.1). Hosts a browser on macOS, Linux and Windows (Windows is a preview: the core host path has run on one real Windows laptop) — see the table below and docs/install.md. Cross-device control works over Tailscale (docs/networking.md); it has been tested with two instances on one machine and with a Mac controlling a Windows host. Any device can also control a host's browser with just the controller client. The in-browser UI and the messages returned to agents are currently in Chinese; English localization is planned.
How it works
① On one computer. Every tool call passes the same checks in the daemon before anything reaches the browser; anything that raises privileges also needs a person to confirm on the host — an agent cannot answer that prompt.
flowchart LR
agent["AI agent<br/>Claude Code, Codex, …"]
subgraph host["Host computer"]
daemon["daemon<br/>🛑 emergency stop<br/>🔑 grant and scope<br/>🌐 site policy<br/>📝 audit log"]
confirm["✋ human confirmation<br/>macOS dialog · Windows Hello · zenity / kdialog"]
helper["UI helper (optional)<br/>Chrome's own bubbles and prompts"]
stub["native messaging host"]
subgraph chrome["Chrome"]
ext["Browser Mesh extension<br/>chrome.debugger"]
tabs["your tabs"]
end
end
agent -- "MCP: stdio, or HTTP on 127.0.0.1" --> daemon
daemon <--> stub <--> ext --> tabs
daemon -- "raising privileges" --> confirm
daemon --> helper -.-> chrome
daemon -. "dedicated profile: start / stop<br/>over --remote-debugging-pipe" .-> chrome② Across your devices. The controlling device needs only the small controller client (or a full install). The host decides what each device may do.
flowchart LR
subgraph laptop["Another device"]
a2["AI agent"] -- "MCP" --> m2["mesh-client"]
end
subgraph home["Host computer"]
d2["daemon<br/>🔍 device = tailscale whois<br/>✋ first connection approved<br/>🔑 grants per device<br/>🛑 emergency stop · 📝 audit"]
c2["Chrome"]
end
m2 -- "Tailscale only, port 47800" --> d2 --> c2③ Remote help. Nothing is installed on the other person's computer except the extension, and the relay sees only ciphertext.
flowchart LR
subgraph you["Your computer"]
a3["your agent"] --> d3["your daemon"]
r3["relay<br/>opened only for the session<br/>forwards ciphertext"]
end
subgraph them["Their computer"]
e3["extension only<br/>✋ check code compared by phone<br/>✋ they click Allow · confirm risky steps<br/>⏱ time limit · 🛑 Stop"]
c3["their Chrome"]
end
d3 -- "end-to-end encrypted" --> r3
e3 -- "end-to-end encrypted" --> r3
e3 --> c3Details: architecture, networking, remote support, security model.
Related MCP server: Omotai Runtime
Platform support and verification
Host | What has been exercised | Limits before release |
macOS | Browser control and the extension, with the screen unlocked and locked; native confirmation; the optional UI helper against Chrome 156's Chinese interface (password bubble, site permissions including location) | The UI helper's English labels have not been seen on macOS (they have on Windows); only the fixed actions listed in installation are implemented |
Linux | Debian container + Xvfb: install/uninstall, daemon, Native Messaging path, confirmation gate, clipboard and screen capture | Chrome with the extension has not been exercised on Linux. Wayland desktop integration is not established by the Xvfb test; X11 privilege changes are refused by default |
Windows | On a real Windows laptop with Chrome: installation (logon task, Native Messaging registration); the Windows Hello confirmation gate; a dedicated profile controlled locally (new tab, navigate, snapshot, find, read text, typing Chinese and clicking, screenshots, emergency stop and lifting it with Windows Hello); control from a Mac through the controller client (list / open / close tabs, read text, screenshots; the first connection approved with Windows Hello); stopping the logon task ( | Preview. Not yet run on a real machine: the UI helper, uninstall, Windows as a controller, your everyday Chrome profile with the extension, clipboard and full-screen capture |
Windows still needs real-machine checks for the items above, plus Task Scheduler restart-on-failure. Windows file access lacks O_NOFOLLOW; its checks are not the same as POSIX. See the Windows checklist. The browser UI helper is implemented on macOS and Windows; Linux AT-SPI is not implemented.
Browser-interface actions require granting Accessibility permission to the separate macOS UI helper. Multi-account autofill selection, keeping a dangerous download, and cancelling the print dialog are not currently provided.
These are implementation and test-status statements, not a completed security audit or a claim that every platform is production-ready.
What it does
MCP server (stdio and Streamable HTTP on
127.0.0.1) exposing 45 tools: tabs, navigation, accessibility snapshots with element refs, screenshots, click / type / fill forms / keys / scroll / hover, JS evaluation, dialogs, console logs, browser lifecycle and windows.Works with your everyday Chrome profile via an MV3 extension + Native Messaging +
chrome.debugger(Chrome 136+ forbids remote-debugging switches on the default profile, so this is the only way), or with a dedicated profile that the daemon launches over--remote-debugging-pipe. Never--remote-debugging-port.You can see and interrupt it: controlled tabs get a yellow frame, an animated virtual cursor and a status label. In yield mode human input pauses the agent; you hand control back with one click (「交还」). In lock mode a page overlay blocks accidental clicks and keys and offers 「接管」 (take over). This is best-effort interaction protection, not a boundary against hostile pages; the popup and CLI also provide emergency stop.
Grants and scopes: tools use scopes (
read,act,storage,browser_admin,settings,system,cdp_raw,browser_ui);batchchecks each step's scope separately. Agents can request more and wait for your approval.Site policies: per-domain
allow/read_only/deny. Password managers, payment dashboards and account-security pages are read-only by default; regional presets (banks, brokers, payment services) are opt-in.Human confirmation gate: lifting an emergency stop, issuing or approving grants, relaxing site rules and enabling lifecycle control of your everyday Chrome require host-side confirmation. macOS uses a native dialog; Windows uses Windows Hello; Linux requires a suitable graphical session. Unsupported confirmation environments refuse by default; see security.md.
Settings and system access (off by default): proxy, privacy, fonts, site permissions and any
chrome://settingspreference with recorded rollback; enterprise policies exported for an admin to install; files in directories you open, clipboard, full-screen capture — see docs/settings.md.Across devices and to anyone: control a Mac's browser from another device on your Tailscale network (docs/networking.md), or help someone outside it who installs only the extension, over an end-to-end encrypted relay you open on demand (docs/remote-support.md).
Emergency stop and audit log: one button detaches everything within milliseconds; every action is recorded in a local SQLite audit log (typed text is stored as length only).
Quick start
Requirements: Google Chrome, Node.js ≥ 22.13, npm, and an interactive desktop session on a supported host. Start with a dedicated profile if you do not want to expose your everyday sign-ins. The commands below install a persistent daemon and register Native Messaging; they are not just a build step.
git clone <this repo> browser-mesh && cd browser-mesh
npm install
scripts/install.sh --chrome # registers the native host + a LaunchAgent for the daemonThe installer above is for macOS. On Linux, use node scripts/install.mjs --chrome after npm install; on Windows (preview), use node scripts\install.mjs --chrome in PowerShell. Instead of --chrome, use --profile-dir DIR to register a non-default Chrome user-data directory. See installation for data locations, service controls and platform dependencies. Keep this checkout in its installed location: the native host and MCP command refer to it.
Then in Chrome: open chrome://extensions, enable Developer mode, click Load unpacked and select packages/extension.
Register the MCP server with your agent, e.g. Claude Code:
claude mcp add browser-mesh -- node "$PWD/packages/daemon/mcp-stdio.mjs"
# or HTTP:
claude mcp add --transport http browser-mesh http://127.0.0.1:47801/mcpCheck it is connected: node packages/daemon/cli.mjs status, or open http://127.0.0.1:47801/status.
Prefer not to let an agent touch your everyday profile? Skip loading the extension there and let the agent call browser_launch — it starts a dedicated profile under ~/Library/Application Support/BrowserMesh/profiles/ (requires the browser_admin scope, which you approve once).
Security model in brief
High privilege, not a sandbox. An everyday browser carries your signed-in sessions. The initial local wildcard grant enables
read+act; choosedefaultGrant: nullbefore the first database is created, or revoke existing grants, if that is too broad. File, settings, clipboard and UI-helper access have additional opt-ins.Trusted host account. Processes with the same OS account can change the code, configuration and grant database. Local agent names are not separate OS identities. The policy/confirmation interfaces do not isolate malicious same-user shell access. Automation with Accessibility permission can also defeat a macOS confirmation dialog.
Limited network exposure. HTTP MCP is loopback-only; the peer listener uses Tailscale. Peer identity is per device, not per person; grant only what you trust processes on that device to do. Remote support additionally requires the humans to compare the check code; encryption alone is not that verification.
Untrusted pages remain untrusted. Site rules and grants reduce permitted operations; the overlay is not an anti-tampering boundary. Treat page text and pushed events as data, never instructions. Emergency stop prevents further dispatch but cannot undo a submission or another action already performed.
Audit data can be sensitive. Text input is summarized rather than logged verbatim, but URLs, account-related metadata and operation summaries can still be private. Keep the runtime data directory out of source archives and public issue attachments.
Read the full security model and limitations before connecting a personal browser. Use a separate profile or machine for higher-risk tasks; this preview does not claim isolation from a compromised host or browser.
Known limitations
Chrome's own interface (
browser_ui, opt-in): handles the "Save password?" bubble and site permission requests. It cannot cancel the print dialog, pick between several saved accounts in the autofill dropdown, or press "Keep" on a dangerous download — see install.md for why.With the macOS screen locked, web-page tools keep working, but
browser_ui, full-screen capture and the confirmation dialog do not — see install.md.Linux has no UI helper yet.
Documentation
Security model and limitations — read this before connecting your everyday profile
Chrome technical notes — what Chrome 136+ allows and forbids for automation
Uninstall
scripts/uninstall.sh # keeps your data (audit log, grants, config)
scripts/uninstall.sh --purge # removes everythingRemove the extension from chrome://extensions.
On Linux or Windows use node scripts/uninstall.mjs (Windows accepts backslashes). --purge also removes runtime data; back up anything you want to keep first.
License
MIT — Copyright (c) 2026 nicgzlfly. Runtime dependencies and their licenses: THIRD_PARTY_NOTICES.md. Report security issues privately — see SECURITY.md. Changes: CHANGELOG.md; contributing: CONTRIBUTING.md.
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
This server cannot be deployed
Maintenance
Related MCP Connectors
Let AI agents use your real, signed-in Chrome: tabs, clicks, forms, screenshots and video.
Control a real browser from AI agents: navigate, click, type, screenshot, read the page.
Let AI agents work in your own signed-in Chrome, with approvals and verified results.
Browser MCP for logged-in tasks. Uses your Chrome — credentials stay local. Zero-token replay.
Related MCP Servers
- AlicenseNot gradedqualityBmaintenanceEnables AI agents to drive a real, already logged-in browser from the CLI or over MCP, with enforceable approval gates for actions and safeguards against prompt injection.2,078,458 npmMozilla Public 2.0

Omotai Runtimeofficial
AlicenseNot gradedqualityBmaintenanceEnables AI agents to safely control a browser through MCP by enforcing deterministic policies, keeping secrets out of model context, and requiring human confirmation for risky actions.Apache 2.0- AlicenseNot gradedqualityCmaintenanceEnables AI clients to control a real, visible Chrome via MCP for navigating, clicking, typing, downloading files, and reading page content. Supports isolated CDP and everyday-Chrome extension modes with user confirmations for risky actions.17 npmMIT

TabBridgeofficial
AlicenseNot gradedqualityBmaintenanceLets any MCP-speaking AI agent drive your everyday Chrome browser, opening pages, reading them, clicking, typing, taking screenshots, and inspecting DevTools console, network, storage and performance data with your existing sign-ins. Risky actions like submitting or paying require your explicit approval, and agents are confined to their own tab group with secrets hidden.Apache 2.0