Skip to main content
Glama
Sign In
neu-browser-mcp

neu-browser-mcp

Official
by neutree-ai
OverviewSchemaSourceRelated ServersDiscussionsAdmin
TypeScript
Remote

Server Configuration

Describes the environment variables required to run the server.

NameRequiredDescriptionDefault
KERNEL_API_KEYNoKernel API key. Required for the `kernel` backend.
NAP_BROWSER_URLNoBase URL of the browser service API. Required for the `nap` backend.
NAP_BROWSER_TOKENNoBearer token for the browser service. Required for the `nap` backend.
BROWSER_USE_API_KEYNoBrowser Use Cloud API key. Required for the `browser-use` backend.
NEU_BROWSER_BACKENDNoPicks the backend and defaults to `nap`. Valid values: `nap`, `kernel`, `browser-use`, `hyperbrowser`.nap
HYPERBROWSER_API_KEYNoHyperbrowser API key. Required for the `hyperbrowser` backend.
NAP_BROWSER_PUBLIC_URLNoBase URL used in file download links. Set it when `NAP_BROWSER_URL` is an internal address users cannot reach. Defaults to `NAP_BROWSER_URL`.

Instructions

Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.

This server publishes no instructions, or was last inspected before Glama recorded them.

Capabilities

Features and capabilities supported by this server

Protocol revision2025-11-25

CapabilityDetails
tools
{
  "listChanged": true
}

Tools

Functions exposed to the LLM to take actions

NameDescription
create_browserA

Create a headless Chrome browser instance. Returns browser_id plus connection info:

  • connect_command: a ready-to-run agent-browser command — just run it as-is to start driving the browser.

  • connect_url: a wss:// URL for "agent-browser connect". Do NOT pass cdp_url here.

  • cdp_url: an https:// URL for Playwright only, via browser.connectOverCDP(cdp_url).

  • live_view_url: share with the user so they can watch. The browser auto-expires after the timeout (default 1 hour, max 24 hours).

list_browsersA

List all active browser instances. Each item includes the same connection info as create_browser (cdp_url, live_view_url, plus connect_url/connect_command for running browsers), so this doubles as a reconnect path — pick a running browser and run its connect_command.

delete_browserB

Stop and destroy a browser instance.

list_browser_filesA

List files inside the browser sandbox. Defaults to /tmp/downloads — point downloads there first, via CDP Browser.setDownloadBehavior with downloadPath: "/tmp/downloads". Use after a download completes to discover what was saved. The browser sandbox is ephemeral, so files only exist while the browser is alive. There can be a brief delay between CDP downloadProgress reporting "completed" and the file being readable; if a file looks truncated, retry after a moment.

get_browser_file_urlA

Get a download URL for a file inside the browser sandbox. Returns the URL alongside the current file state — use ready: true to confirm the file exists. The URL embeds an auth token and can be handed to the user (e.g. as a clickable link in chat) or fetched directly. Files are only available while the browser is alive.

Prompts

Interactive templates invoked by user choice

NameDescription

No prompts

Resources

Contextual data attached and managed by the client

NameDescription

No resources

Deploy Server
Let your AI agent see the webWeb-to-screenshot MCP for Claude, Cursor and Codex. No cookie banners. Free to start.snap-render.comAd

TDQS

A3.9/5.0

Scored across 5 tools

Disambiguation4/5

Each tool has a distinct role: lifecycle (create/list/delete), file listing, and file URL retrieval. The only mild overlap is that list_browsers re-exposes connection info also returned by create_browser, but that is a documented reconnect convenience rather than a genuine ambiguity.

Naming Consistency5/5

All five tools follow a consistent snake_case verb_noun pattern (create_browser, list_browsers, delete_browser, list_browser_files, get_browser_file_url). The singular/plural browser vs. browsers variation is trivial and matches the resource semantics.

Tool Count5/5

Five tools is well-scoped for a browser-sandbox lifecycle server: create, list, delete, and two file-access tools. Nothing feels padded or missing to the point of being thin.

Completeness4/5

Full lifecycle (create/list/delete) plus sandbox file discovery and download URL generation are covered, which is appropriate given that actual browsing is delegated to the external agent-browser/CDP connection. Minor gaps exist, such as no way to extend a browser's timeout or delete sandbox files.

Maintenance

ActivityMaintained
ResponsivenessNo issues

Resources