chrome-orchestrator

chrome-orchestrator

Enables AI agents to dynamically provision headless browser environments and connect to them through a unified HTTP/SSE MCP endpoint with persistent storage and automatic health monitoring.

Category
访问服务器

README

Chrome Orchestrator

chrome-orchestrator is a lightweight Python SDK and CLI that abstracts away Docker management for headless browser environments, specifically Arthur and gloria-headless.

It runs a long-lived background daemon that acts as a unified HTTP proxy. AI agents request a browser session, and the daemon dynamically provisions the requested Docker container, mounts persistent storage, and seamlessly proxies the stateless HTTP/SSE MCP (Model Context Protocol) endpoints on port 3000.

Features

  • Daemon Proxy Architecture: Exposes a single, unified port (3000 by default) and proxies MCP traffic to dynamically allocated Docker containers (port 8000 for Arthur, 8787 for Gloria).
  • Agent Manifests: The SDK returns a ContainerManifest containing environmental instructions so AI agents automatically understand the capabilities of the environment they just provisioned.
  • Persistent Storage: Automatically mounts a local directory (~/.chrome-orchestrator/data/<image>) to persist state across browser sessions.
  • Active Health Monitoring: The daemon actively monitors container health in the background and attempts restarts if they become unresponsive.

Environments

Both environments expose the identical execute_python tool using the chrome browser object, allowing AI agents to seamlessly write Python code regardless of the underlying backend.

  • Arthur (Default): A pure, minimal Chromium environment with exact DOM fidelity, interacting natively via CDP WebSockets. Optimal for token limits and standard browsing.
  • Gloria-Headless: An environment pre-configured with uBlock Origin Lite that interacts via the Chrome Bridge extension (Native Messaging). Use this for "hard stuff" like captcha evasion, since it behaves more like a real user browser.

Requirements

  • Python 3.9+
  • Docker or Podman installed and running. For Docker, ensure permissions to access the Docker daemon (/var/run/docker.sock). For Podman, rootful socket access is supported natively (unix:///run/podman/podman.sock) as a seamless fallback.

Installation

# From PyPI
pip install chrome-orchestrator

# Or without installing via uvx
uvx --from chrome-orchestrator orchestrator <command>

# Or from source
pip install -e .

Usage

1. Start the Daemon

First, start the background daemon proxy. This service must be running for the CLI and SDK to work.

uvx --from chrome-orchestrator orchestrator daemon

By default, the daemon runs on http://0.0.0.0:3000.

2. CLI Usage

You can manage arthur and gloria containers via the CLI:

# Provision an Arthur container
uvx --from chrome-orchestrator orchestrator arthur up

# Stop an Arthur container
uvx --from chrome-orchestrator orchestrator arthur down

# Provision a Gloria-headless container
uvx --from chrome-orchestrator orchestrator gloria up

# Stop all running orchestrator containers
uvx --from chrome-orchestrator orchestrator down

When you run up, the orchestrator will provision the container and output a static MCP HTTP URL you can connect your client to, such as http://localhost:3000/arthur/mcp.

3. Python SDK Usage

AI agents and Python scripts can use the SDK to programmatically provision and tear down environments:

from chrome_orchestrator import sdk

# Provision the environment (defaults to Arthur)
manifest = sdk.up(sdk.ContainerImage.ARTHUR)

print(f"Connected! MCP URL: {manifest.mcp_url}")
print(f"Agent Instructions:\n{manifest.instructions}")

# Your automation/MCP client connects to `manifest.mcp_url` here...

# Tear down the session when finished
sdk.down(sdk.ContainerImage.ARTHUR)

Testing

The project uses pytest for testing. The tests target the highest seams (CLI and SDK) and simulate HTTP proxy interactions.

pytest tests/

推荐服务器

Baidu Map

Baidu Map

百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。

官方
精选
JavaScript
Playwright MCP Server

Playwright MCP Server

一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。

官方
精选
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。

官方
精选
本地
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。

官方
精选
本地
TypeScript
VeyraX

VeyraX

一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。

官方
精选
本地
graphlit-mcp-server

graphlit-mcp-server

模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。

官方
精选
TypeScript
Kagi MCP Server

Kagi MCP Server

一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。

官方
精选
Python
e2b-mcp-server

e2b-mcp-server

使用 MCP 通过 e2b 运行代码。

官方
精选
Neon MCP Server

Neon MCP Server

用于与 Neon 管理 API 和数据库交互的 MCP 服务器

官方
精选
Exa MCP Server

Exa MCP Server

模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。

官方
精选