browser-mcp-claude
Click on "Install Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@browser-mcp-claudeOpen example.com and show me the page content"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Browser MCP for Claude and other models
This is a remote MCP server that provides general browser tools via Playwright and Streamable HTTP. It does not rely on Claude inside the server; any client that supports MCP over HTTPS can use it, including Claude, or a middleware layer using Gemini, OpenAI, or a local model.
The server opens and reads pages and fills non-sensitive fields, but it blocks password fields, files, final submission, and payment or deletion operations.
Deploying on Replit
Import this repository into Replit using Docker, then choose a Reserved VM or an always-on service. Add the following variables in Secrets:
Key | Value |
| A long random token to authenticate Claude with the server |
|
|
|
|
Do not put passwords, verification codes, or site keys inside GitHub or the code.
After deployment, the endpoint becomes:
https://YOUR-REPL-DOMAIN/mcp
It must be public over HTTPS; Claude connects to the server from Anthropic's cloud infrastructure, not from your local phone.
Related MCP server: mcp-browser
Connecting Claude Web
In Claude on the phone, open Customize → Connectors → + → Add custom connector, and enter the /mcp link. If the interface asks for OAuth Client ID/Secret, do not put MCP_AUTH_TOKEN in those fields; those fields are for OAuth. Use the authentication mechanism that Claude offers for the account, or implement the connection via the Claude API if the server needs a direct Bearer token.
After adding, enable the connector in the conversation from + → Connectors. Start with read-only commands like opening a public page and taking a text snapshot. Review every tool request before allowing it.
Multi-model
The MCP server does not enforce a model. Claude, Gemini, OpenAI, or any other model that supports MCP is the client that decides to call the tools. The model can be changed without modifying Playwright or the MCP server.
Server tools
browser_open(url): Open a public HTTP(S) page.browser_snapshot(): Read the current page and display references to interactive elements.browser_click(ref): Click a non-final element.browser_fill(ref, value): Fill a non-sensitive text field.browser_select(ref, label_or_value): Select an item from a list.browser_back(): Go back one page.browser_close(): Close the browser session.
Security limitations
The server refuses to run if MCP_AUTH_TOKEN is not set. It must run behind public HTTPS, and access to sites and data should be restricted as needed. Do not use it to bypass CAPTCHA or site controls, and do not give Claude automatic permission for purchase, deletion, or submission operations. Add real OAuth before using it with production accounts.
References
This server cannot be installed
Maintenance
Resources
Unclaimed servers have limited discoverability.
Looking for Admin?
If you are the server author, to access and configure the admin panel.
Related MCP Servers
- FlicenseBqualityDmaintenancePlaywright wrapper for MCP that enables LLM-powered clients to control a browser for automation tasks.101
- FlicenseNot gradedqualityCmaintenanceExposes Playwright browser automation as MCP tools, enabling AI assistants to control a real browser tab-by-tab for form filling, navigation, and more, while preserving the user's active session.
- FlicenseNot gradedqualityCmaintenanceEnables browser automation through the MCP protocol, allowing AI agents to control a real browser using accessibility snapshots and natural language commands.
- AlicenseNot gradedqualityBmaintenanceEnables autonomous web QA by exposing Playwright browser control as MCP tools for navigation, accessibility snapshotting, interaction, and bug detection.3MIT
Related MCP Connectors
Stealth web browser for agents: search, fetch, click and type through persistent sessions over MCP.
A paid remote MCP for AI agent browser DevTools MCP, built to return verdicts, receipts, usage logs,
Provides cloud browser automation capabilities using Stagehand and Browserbase, enabling LLMs to i…
Latest Blog Posts
- Who's Calling? MCP Hosts Are an Identity Blind Spot (And the Spec Knows It)By Om-Shree-0709 on .mcpAgent IdentityOAuth 2.1
- Your AI Chatbot Just Exposed Your CEO's Salary to an InternBy Om-Shree-0709 on .Agent IdentityMCP SecurityOAuth Delegation
- Why MCP Servers Need Execution Sandboxing (And Why Your Current Stack Isn't Enough)By Om-Shree-0709 on .Agentic AiPrompt InjectionWebAssembly
MCP directory API
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/saGoldberg31/browser-mcp-claude'
If you have feedback or need assistance with the MCP directory API, please join our Discord server