How do I install openbrowser-ai?

Install via Claude Code (local): 1. Install Claude Code CLI 2. Copy the command below, replace with real env values, then run in terminal 3. See Environment variables below if listed Example: claude mcp add openbrowser-ai -- uvx openbrowser-ai

What is openbrowser-ai good for?

openbrowser-ai is a fit for: Browser automation. See AgentHub scenario pages for pairings.

openbrowser-ai MCP Setup & Guide

MCP ServerMCP RegistryOfficial

me.openbrowser/openbrowser-ai · v0.1.54

AI browser automation. Write async Python to navigate, click, type, and extract data.

Repository ↗

Copy the install config on this page first, then verify docs and permissions upstream.

Overview

AI browser automation. Write async Python to navigate, click, type, and extract data. openbrowser-ai is a MCP Server listed from MCP Registry. Transports: stdio. This page includes an overview, setup tutorial, install commands, and use cases for Trae, Tongyi Lingma, Cursor, Claude Code, and VS Code.

Use cases

Browser automation

AgentHub Verified AvailabilityTested & Ready

Automated pipeline validated install commands, protocol & client compatibility

Verified At2026-10-04

Install Snippets TestedCLI & JSON config syntax verified

Protocol Handshake ReadyComplies with JSON-RPC 2.0 specifications

Origin Registry ActiveSourced from official-mcp-registry

Verified ClientsClaude Code、Claude Desktop、Cursor 等

Security Tier: A+ 级 · 官方认证推荐 (A+)·Maintained by official/verified teams, audited for standard MCP protocol compliance.

Score 90/100

Copy by platform

Choose your platform

Open or create .cursor/mcp.json in your project root
Click Copy config and paste; merge only this mcpServers entry if others exist
Replace <placeholders> in env with real secrets (see Environment variables below)
Save, then Cmd+Shift+P → Reload Window

Pre-fill Environment Variables (Optional)

OPENBROWSER_HEADLESS

OPENBROWSER_ALLOWED_DOMAINS

OPENBROWSER_USER_DATA_DIR

OPENBROWSER_STORAGE_STATE

OPENBROWSER_COMPACT_DESCRIPTION

OPENBROWSER_MAX_OUTPUT

ANONYMIZED_TELEMETRY

Privacy Guarantee: Keys are replaced 100% locally in your browser memory and are NEVER uploaded or stored on any server.

Click to copy snippet

{
  "mcpServers": {
    "openbrowser-ai": {
      "command": "uvx",
      "args": [
        "openbrowser-ai"
      ]
    }
  }
}

Setup tutorial

Open the openbrowser-ai page and confirm this MCP Server (source: MCP Registry).
Copy the Cursor, Claude Code, or VS Code snippet.
Merge it into mcpServers and replace env placeholders with real secrets.
Reload the window, then call the MCP tools from your agent chat.

Install commands

Install commands and setup steps are in the HTML so search engines and no-JS browsers can read them without running client JavaScript.

Claude Code (local)

Install Claude Code CLI
Copy the command below, replace <placeholders> with real env values, then run in terminal
See Environment variables below if listed

claude mcp add openbrowser-ai -- uvx openbrowser-ai

Cursor — .cursor/mcp.json (local)

Open or create .cursor/mcp.json in your project root
Click Copy config and paste; merge only this mcpServers entry if others exist
Replace <placeholders> in env with real secrets (see Environment variables below)
Save, then Cmd+Shift+P → Reload Window

{
  "mcpServers": {
    "openbrowser-ai": {
      "command": "uvx",
      "args": [
        "openbrowser-ai"
      ]
    }
  }
}

VS Code — .vscode/mcp.json (local)

Install GitHub Copilot in VS Code with MCP support
Open or create .vscode/mcp.json in your project root
Click Copy config and paste; merge only this mcpServers entry if others exist
Replace <placeholders> in env with real secrets
Save and Developer: Reload Window

{
  "mcpServers": {
    "openbrowser-ai": {
      "command": "uvx",
      "args": [
        "openbrowser-ai"
      ]
    }
  }
}

Claude Desktop — claude_desktop_config.json (local)

Open Claude Desktop claude_desktop_config.json (see remote guide for paths)
Click Copy config and merge under mcpServers
Replace <placeholders> in env with real secrets
Fully quit and restart Claude Desktop

{
  "mcpServers": {
    "openbrowser-ai": {
      "command": "uvx",
      "args": [
        "openbrowser-ai"
      ]
    }
  }
}

Trae — .trae/mcp.json (local)

Trae → Settings → MCP, or edit .trae/mcp.json / global mcp.json
Click Copy config and merge mcpServers
Replace <placeholders> in env with real secrets
Save and reload Trae

{
  "mcpServers": {
    "openbrowser-ai": {
      "command": "uvx",
      "args": [
        "openbrowser-ai"
      ]
    }
  }
}

Cherry Studio — MCP settings (local)

Cherry Studio → Settings → MCP Servers → Add (STDIO)
Or import JSON: click Copy config and merge mcpServers
Replace <placeholders> in env; ensure Node.js / uv (npx, uvx) are installed
Enable the server and check tools load

{
  "mcpServers": {
    "openbrowser-ai": {
      "command": "uvx",
      "args": [
        "openbrowser-ai"
      ]
    }
  }
}

Tongyi Lingma — MCP config (local)

Lingma Settings → MCP → + → STDIO or config file
Click Copy config and merge mcpServers
Replace <placeholders> in env; need Node.js 18+ (npx) or uv (uvx)
Confirm connected before using tools in agent chat

{
  "mcpServers": {
    "openbrowser-ai": {
      "command": "uvx",
      "args": [
        "openbrowser-ai"
      ]
    }
  }
}

Windsurf — mcp_config.json (local)

Edit ~/.codeium/windsurf/mcp_config.json
Click Copy config and merge mcpServers (stdio same as Cursor)
Replace <placeholders> in env, save, refresh Cascade

{
  "mcpServers": {
    "openbrowser-ai": {
      "command": "uvx",
      "args": [
        "openbrowser-ai"
      ]
    }
  }
}

Cline — MCP Servers (local)

Cline panel → Settings → MCP Servers
Click Copy config and merge
Replace <placeholders> in env, then save

{
  "mcpServers": {
    "openbrowser-ai": {
      "command": "uvx",
      "args": [
        "openbrowser-ai"
      ]
    }
  }
}

WorkBuddy — .workbuddy/mcp.json (local)

Edit ~/.workbuddy/mcp.json (user) or project .workbuddy/mcp.json
Or Plugins → MCP Servers → Configure MCP and paste the JSON below
Replace <placeholders> in env; on Windows prefer absolute paths for command/scripts
Save, restart WorkBuddy, confirm connector status is green

{
  "mcpServers": {
    "openbrowser-ai": {
      "command": "uvx",
      "args": [
        "openbrowser-ai"
      ]
    }
  }
}

Troubleshooting & Common ErrorsFAQ

Common connection errors and verified fixes for openbrowser-ai

Getting 'connection closed' or exit code 1 in Cursor / Claude Code for openbrowser-ai?

Usually caused by missing runtime paths or IDE environment inheritance. Fix steps: 1. Verify Node.js 18+ (npx) or Python 3.10+ (uvx) is installed; 2. Run 'which npx' or 'which uvx' in terminal, and replace 'command' with absolute path; 3. Reload or restart the client window after modifying config.

Fix Snippet

# Check binary path in terminal:
which npx
node -v

Getting 'spawn npx ENOENT' or 'command not found'?

The editor background process does not inherit your full terminal PATH. Solution: Globally install the package, or set the absolute binary path (e.g. C:\Program Files\nodejs\npx.cmd on Windows, or /usr/local/bin/npx on macOS).

Missing required environment variable or authentication failure?

openbrowser-ai requires environment variables (OPENBROWSER_HEADLESS、OPENBROWSER_ALLOWED_DOMAINS、OPENBROWSER_USER_DATA_DIR、OPENBROWSER_STORAGE_STATE、OPENBROWSER_COMPACT_DESCRIPTION、OPENBROWSER_MAX_OUTPUT、ANONYMIZED_TELEMETRY). Ensure you have added valid keys under the 'env' object in your config file without trailing whitespace.

Fix Snippet

// .cursor/mcp.json 或 claude_desktop_config.json
{
  "env": {
    "API_KEY": "your_actual_key_here"
  }
}

Still not connecting? Read the MCP troubleshooting guide →

Tool Mock Playground

Sandbox

openbrowser-ai Tool Interface Simulation · Preview tool schema & outputs without local runtime

Sandbox Ready

fn: openbrowser_aiExecute the core tool interface of openbrowser-ai

Request ArgumentsJSON Schema

{
  "target": "openbrowser-ai",
  "action": "execute",
  "options": {
    "mode": "standard",
    "timeoutMs": 5000
  }
}

💡Parameters generated dynamically by Agent runtime

Agent Tool Output

Click 'Run Mock' above

to preview the raw response returned to the LLM

Env: AgentHub Virtual SandboxJSON-RPC 2.0

Decision Guide: Why & When to Use

Assess suitability before installing to save trial-and-error time

Best Suited For

E2E testing
Headless screenshots
Form automation

When NOT to Use

Bulk fetch when a stable REST API exists
CAPTCHA bypass (compliance risk)

v1.2.25

SkillClawHub87.2k

io.clawhub.byungkyu/api-gateway

Call third-party APIs through the Maton gateway, which injects the credential for an app the user has already connected. Use this skill when the user names a connected app and a concrete action in it - read a mailbox, query a CRM, file an issue, update a spreadsheet, etc.

source

View on ClawHub

v1.2.6

SkillClawHub13.5k

io.clawhub.byungkyu/linkedin-api

LinkedIn API integration with managed OAuth. Share posts, manage profile, and access LinkedIn features. Use this skill when users want to share content on LinkedIn, get profile/organization information, or interact with LinkedIn's platform.

source

View on ClawHub

WhatsApp Business

v1.2.8

SkillClawHub24.1k

io.clawhub.byungkyu/whatsapp-business

WhatsApp Business API integration with managed OAuth. Send messages, manage templates, and handle conversations. Use this skill when users want to interact with WhatsApp Business. For other third party apps, use the api-gateway skill (https://clawhub.ai/byungkyu/api-gateway).

source

View on ClawHub

Salesforce

v1.2.7

SkillClawHub20.1k

io.clawhub.byungkyu/salesforce-api

Salesforce CRM API integration with managed OAuth. Install only if you need Salesforce CRM administration. Connect with the narrowest Salesforce permissions available, prefer sandbox orgs for destructive or batch work, verify the intended connection ID before each request, etc.

source

View on ClawHub

Keep exploring AgentHub

Most people compare similar tools or check scenario guides before installing—start here.

Browse more like this Search similar Pick by scenario Prompt playbooks MCP hands-on Reviews & comparisons

Listing badge: put AgentHub on your site

Copy either snippet into your project homepage, docs, or GitHub README. The badge is a hotlinked SVG — nothing to host — and it links back to this page so visitors can find the install steps.

Preview

HTML

<a href="https://myagenthub.cn/p/me.openbrowser/openbrowser-ai" title="Listed on AgentHub: openbrowser-ai" target="_blank" rel="noopener">
  <img src="https://myagenthub.cn/badge/me.openbrowser/openbrowser-ai?lang=en" alt="Listed on AgentHub: openbrowser-ai" height="20" style="border:0"/>
</a>

Markdown (GitHub README)

[![Listed on AgentHub: openbrowser-ai](https://myagenthub.cn/badge/me.openbrowser/openbrowser-ai?lang=en)](https://myagenthub.cn/p/me.openbrowser/openbrowser-ai)

Badges are generated on the fly from /badge/<package-id>, so name and listing changes propagate automatically. Keep the link target unchanged — it is what counts as the referral.

Unified Manifest

{
  "id": "me.openbrowser/openbrowser-ai",
  "type": "mcp-server",
  "version": "0.1.54",
  "displayName": "openbrowser-ai",
  "description": "AI browser automation. Write async Python to navigate, click, type, and extract data.",
  "repository": {
    "url": "https://github.com/billy-enrizky/openbrowser-ai.git",
    "source": "github"
  },
  "distribution": {
    "packages": [
      {
        "registryType": "pypi",
        "identifier": "openbrowser-ai",
        "version": "0.1.54",
        "runtimeHint": "uvx",
        "transport": "stdio",
        "environmentVariables": [
          {
            "name": "OPENBROWSER_HEADLESS",
            "description": "Run browser in headless mode (true/false)"
          },
          {
            "name": "OPENBROWSER_ALLOWED_DOMAINS",
            "description": "Comma-separated list of allowed domains (empty for no restrictions)"
          },
          {
            "name": "OPENBROWSER_USER_DATA_DIR",
            "description": "Optional explicit Chrome profile directory. Omit for an isolated profile per MCP server instance; concurrent owners of one explicit profile are rejected"
          },
          {
            "name": "OPENBROWSER_STORAGE_STATE",
            "description": "Optional JSON file used as a one-time seed for cookies plus localStorage in managed MCP profiles; later updates are saved to the active profile copy"
          },
          {
            "name": "OPENBROWSER_COMPACT_DESCRIPTION",
            "description": "Use minimal tool description (~500 tokens instead of ~1,430)"
          },
          {
            "name": "OPENBROWSER_MAX_OUTPUT",
            "description": "Maximum output characters per code execution (default: 10000)"
          },
          {
            "name": "ANONYMIZED_TELEMETRY",
            "description": "Enable anonymized usage telemetry (true/false)"
          }
        ]
      }
    ],
    "remotes": []
  },
  "dependencies": [],
  "installTargets": [
    "claude-code",
    "claude-desktop",
    "cursor",
    "vscode",
    "trae",
    "cherry-studio",
    "lingma",
    "windsurf",
    "cline",
    "workbuddy"
  ],
  "keywords": [],
  "provenance": {
    "origin": "official-mcp-registry",
    "originalId": "me.openbrowser/openbrowser-ai",
    "originalUrl": "https://registry.modelcontextprotocol.io/v0.1/servers/me.openbrowser%2Fopenbrowser-ai/versions/latest",
    "isOfficial": true,
    "status": "active"
  }
}