AgentRPA
Enables browser-based e-commerce information collection on Taobao pages, including opening search pages and extracting titles and prices with page snapshots under policy controls.
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.
rpa-llm-bridge
把 LLM 接到 RPA 上:一套统一的 action 契约,同时驱动浏览器、Windows 桌面应用和 Office 文件;用 YAML 描述流程,用策略闸门管住不可逆动作,用审计日志留下每一步的痕迹。
内核代号 AgentRPA,Python 包名 agentrpa。
通过 MCP 暴露给 Codex / Claude Desktop / Cursor 等客户端,让模型能真正"操作"本机软件,而不是只会说。
这个项目和"电商采集器"有什么不同
多数 RPA 项目只覆盖一个网站、一种交互。AgentRPA 把差异抽到三层里:
维度 | 单站点采集器 | AgentRPA |
作用域 | 一个网站 | 浏览器 / 桌面应用 / Office 文件 |
流程定义 | 写死在代码里 | YAML Recipe,可 diff、可版本化、可参数化 |
安全边界 | 分散在各函数里 | 集中式 |
不可逆动作 | 靠自觉 | 默认拦截 + 人工确认 + 断点续跑 |
可追溯性 | 打日志了事 | 结构化审计 jsonl,敏感字段强制脱敏 |
失败恢复 | 重跑整个脚本 | run 状态落盘, |
Related MCP server: fast-hands
架构

三个核心概念:
Driver:实现
BaseDriver.execute(action, params)。新增一种被控端 = 新增一个驱动,执行器与策略层零改动。Recipe:YAML 声明的步骤序列。步骤可声明
hitl: true强制人工确认,可声明on_error: skip|retry|abort。Policy:唯一的安全判定点。域名白名单、应用白名单、敏感字段拦截、状态变更闸门都在这里,驱动不许自己另立标准。
快速开始
要求 Windows 10/11(桌面驱动)/ macOS·Linux(浏览器与 Office 驱动)、Python 3.11+。
git clone https://github.com/c8xm/rpa-llm-bridge.git
cd rpa-llm-bridge
python -m venv .venv
.\.venv\Scripts\python.exe -m pip install -U pip
.\.venv\Scripts\python.exe -m pip install -e ".[dev]"
.\.venv\Scripts\python.exe -m playwright install chromium
Copy-Item .env.example .env启动(stdio 模式会被客户端拉起,直接运行会停在等待状态):
.\.venv\Scripts\python.exe server.pyMCP 客户端配置
Codex(%USERPROFILE%\.codex\config.toml):
[mcp_servers.agentrpa]
command = "<PROJECT_DIR>/.venv/Scripts/python.exe"
args = ["<PROJECT_DIR>/server.py"]
cwd = "<PROJECT_DIR>"
enabled = true
startup_timeout_sec = 30
tool_timeout_sec = 180
[mcp_servers.agentrpa.env]
PLAYWRIGHT_HEADLESS = "false"
ALLOW_STATE_CHANGING_ACTIONS = "false"
ALLOWED_DOMAINS = "jd.com,taobao.com,tmall.com"Cursor / Claude Desktop(JSON):
{
"mcpServers": {
"agentrpa": {
"command": "<PROJECT_DIR>\\.venv\\Scripts\\python.exe",
"args": ["<PROJECT_DIR>\\server.py"],
"env": {
"PLAYWRIGHT_HEADLESS": "false",
"ALLOW_STATE_CHANGING_ACTIONS": "false",
"ALLOWED_DOMAINS": "jd.com,taobao.com,tmall.com"
}
}
}
}内置流程
Recipe | 场景 | 驱动 | 说明 |
| 电商信息采集 | browser | 打开搜索页,抽取标题与价格,附页面快照 |
| 通用网页 RPA | browser | 填普通字段,提交前强制人工确认 |
| 桌面 ERP 录入 | desktop | 启动客户端、定位窗口、填单据、读回显 |
| 企业报表 | office | 读源表、写汇总表 |
| 内容平台运营 | browser | 填草稿、截图预览,发布动作人工确认 |
自定义流程只要往 agentrpa/recipes/examples/ 丢一个 YAML:
name: my_flow
description: 我的流程
driver: browser
vars:
target: "https://example.com"
steps:
- id: open
action: navigate
url: "{{target}}"
- id: grab
action: extract
multi:
titles: "h2"
limit: 10工具清单
工具 | 作用 |
| 版本、驱动能力、策略摘要、目录 |
| 白名单与拦截规则 |
| 三个驱动的健康检查 |
| 浏览器生命周期 |
| 打开白名单域名 URL |
| 标题、正文摘要、链接、表单、meta |
| 列出可交互元素并分配短期 ref |
| 点击元素(命中交易类文案会被拦截) |
| 写普通输入框(敏感字段拒绝) |
| 滚动、截图、按选择器批量抽取 |
| 流程发现与执行 |
| 断点查看与人工批准后继续 |
| 审计日志(已脱敏) |
| 桌面窗口与控件 |
| Excel / Word 读写 |
安全模型
域名白名单:
ALLOWED_DOMAINS之外一律拒绝,含同形域名(jd.com.evil.cn)。应用白名单:桌面驱动在
ALLOWED_APPS为空时拒绝启动任何进程。敏感字段:密码、验证码、OTP、银行卡、CVV、身份证等标签直接拒绝输入,审计日志同步脱敏。
状态变更闸门:购买、支付、下单、发布、发送、删除等动作默认拦截;需要人工确认时流程停在
awaiting_human,由人决定是否run_resume。不注入、不伪造:不往页面里注入反检测脚本,不伪造浏览器指纹,不绕过验证码与风控。
可追溯:每一步的 action、判定结果、耗时都落到
artifacts/audit/<date>.jsonl。
项目结构
rpa-llm-bridge/
├─ server.py MCP 工具入口
├─ agentrpa/
│ ├─ config.py 环境变量配置
│ ├─ core/
│ │ ├─ types.py Step / StepResult / RunResult / Decision
│ │ ├─ policy.py 白名单与闸门
│ │ ├─ runner.py 执行器与断点续跑
│ │ ├─ audit.py 审计日志与脱敏
│ │ └─ registry.py 驱动注册表
│ ├─ drivers/
│ │ ├─ base.py 统一 action 契约
│ │ ├─ web.py Playwright 驱动
│ │ ├─ desktop.py pywinauto 驱动
│ │ └─ office.py Excel / Word 驱动
│ └─ recipes/
│ ├─ loader.py YAML 校验与模板渲染
│ └─ examples/*.yaml 五个场景示例
├─ tests/ 43 个单元测试,不需要打开浏览器
├─ docs/DESIGN.md 设计取舍
└─ .github/workflows/ci.yml 多版本 CI开发
.\.venv\Scripts\python.exe -m pytest
.\.venv\Scripts\python.exe -m pytest tests/test_runner.py -v单元测试只覆盖纯逻辑(策略、执行器、模板、审计),不启动浏览器与桌面进程,因此 CI 在 Linux 上也能跑。

图片由 scripts/make_assets.py 调用真实命令生成,不是手绘示意图。
路线图
Recipe 步骤级重试策略可视化(当前只有计数)
多驱动混合流程(浏览器取数 → Excel 落表 → 桌面录入)
录制回放:把一次人工操作转成 Recipe 草稿
审计日志按 run 聚合查询与导出
桌面驱动的控件级截图对比回归
上游与许可
MIT。本项目在架构思路上参考了 JD-Taobao-MCP 的人机协同与输出契约设计,代码为独立实现,来源与改动记录见 UPSTREAM.md。
This server cannot be deployed
Maintenance
Related MCP Connectors
Operate Linux, macOS and Windows from your LLM. Every action runs through an auditable allowlist.
Build and run agents and automations across hundreds of apps using natural language.
Connect AI assistants to your approved computer for files, commands, browser and desktop tasks.
Give AI agents 30,000+ safe, token-optimized actions across Workday, SAP, Oracle + hundreds more.
Related MCP Servers
- AlicenseNot gradedqualityBmaintenanceEnables AI agents to securely control a user's existing Chrome profile locally, providing typed browser actions, form and editor support, WordPress workflows, terminal automation, and Figma inspection with policy-based authorization and redacted auditing.MIT
- AlicenseNot gradedqualityBmaintenanceEnables AI agents to execute local commands and multi-step workflows with durable checkpoints, operator controls, and safe resume/revise capabilities. It also provides local web and YouTube research tools, with optional Windows UI automation.12 npmMIT
- AlicenseNot gradedqualityBmaintenanceEnables AI agents to observe and control the user's live desktop by listing apps, reading accessibility trees, and performing clicks, typing, scrolling, dragging, and other input actions with per-action approvals and local audit archives.1MIT
- AlicenseNot gradedqualityCmaintenanceEnables AI assistants to drive Windows desktop automation — opening and interacting with applications, native UI controls, keyboard/mouse input, files, processes, clipboard, and Chrome/Edge browsing via UI Automation and CDP. All actions pass through a safety policy with human approval, emergency stop, and deterministic verification.MIT