Server facts
Browserbase MCP Server
Project
- Author
- browserbase
- Website
- stagehand.dev
- License
- Apache-2.0
- Category
- Browser Automation
Tech
- Runtime / language
- node
- Transports
- stdio
- Remote available
- No
- Authentication
- Not specified
Activity
- GitHub stars
- 3,399Archived on GitHub
- Last commit
- Data as of
Tags
- Browser Automation
- Browserbase
- Cloud Browser
- Data Extraction
- Stagehand
Overview
Browserbase MCP Server connects AI assistants to cloud browser automation using Browserbase and Stagehand. Its interface is deliberately compact: six tools start or reuse a session, close it, navigate to an address, perform a page action, identify actionable elements, and extract information. Actions and observations accept instructions, making the server a bridge between conversational requests and browser activity. The repository is a self-hostable counterpart to Browserbase's hosted MCP service and describes the tools and functionality as matching. It is aimed at workflows that need both interaction with live pages and extraction of their contents.
The README recommends the hosted service over Streamable HTTP as the simplest starting point. Clients without native support can connect through mcp-remote. For self-hosted stdio use, the npm configuration requires a Browserbase API key, project ID, and a Gemini key for the default model. Source and Docker installation routes are also shown. Self-hosted options control proxies, retained sessions, persistent contexts, viewport dimensions, and model selection. Alternative Stagehand-supported models require the caller's own provider key. Verified Identity is limited to Scale plan users. The hosted route includes Gemini model costs according to the README, whereas self-hosted model customization is explicitly configured by the operator. Those differences help determine whether minimal setup or local configuration flexibility is the better fit.
Install & configure
Requirements
- An MCP-compatible client
- npx for local npm execution
- Browserbase credentials and a model key for self-hosted execution
npx @browserbasehq/mcpThe command is the self-hosted stdio option. The remote URL uses Streamable HTTP and is the README's recommended setup. Self-hosted custom models use --modelName and --modelApiKey.
Examples are translated from the published configuration. Supply your own credentials and verify requirements in the upstream README. Remote connectors are shown only where a supported configuration format is known.
Config file · macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
Windows: %APPDATA%\Claude\claude_desktop_config.json
Use Settings → Developer → Edit Config when available. Do not assume a Linux configuration path; check the client documentation for your platform. Remote servers use the app’s connector flow when supported, not the local stdio file shown here. Merge this example with existing settings; do not overwrite unrelated servers.
{
"mcpServers": {
"mcp-server-browserbase": {
"command": "npx",
"args": [
"@browserbasehq/mcp"
],
"env": {
"BROWSERBASE_API_KEY": "<your Browserbase API key>",
"BROWSERBASE_PROJECT_ID": "<your Browserbase project ID>",
"GEMINI_API_KEY": "<your Gemini API key for the default model>"
}
}
}
}Config file · Project: .cursor/mcp.json
User: ~/.cursor/mcp.json
The project path is relative to the workspace root. The tilde denotes your user home directory. Use Cursor’s MCP settings to confirm which configuration and scope your installed version is reading. Merge this example with existing settings; do not overwrite unrelated servers.
{
"mcpServers": {
"mcp-server-browserbase": {
"command": "npx",
"args": [
"@browserbasehq/mcp"
],
"env": {
"BROWSERBASE_API_KEY": "<your Browserbase API key>",
"BROWSERBASE_PROJECT_ID": "<your Browserbase project ID>",
"GEMINI_API_KEY": "<your Gemini API key for the default model>"
}
}
}
}Config file · Workspace: .vscode/mcp.json
For user scope, use the “MCP: Open User Configuration” command if available. Do not guess a user configuration path across operating systems, profiles, containers, or remote workspaces; check the client documentation. Merge this example with existing settings; do not overwrite unrelated servers.
{
"servers": {
"mcp-server-browserbase": {
"type": "stdio",
"command": "npx",
"args": [
"@browserbasehq/mcp"
],
"env": {
"BROWSERBASE_API_KEY": "<your Browserbase API key>",
"BROWSERBASE_PROJECT_ID": "<your Browserbase project ID>",
"GEMINI_API_KEY": "<your Gemini API key for the default model>"
}
}
}
}Config file · macOS/Linux: ~/.codeium/windsurf/mcp_config.json
Windows: %USERPROFILE%\.codeium\windsurf\mcp_config.json
Open the MCP configuration through Windsurf’s settings when possible to confirm the active file. Paths can differ in managed or future releases; check the client documentation if this location is not present. Merge this example with existing settings; do not overwrite unrelated servers.
{
"mcpServers": {
"mcp-server-browserbase": {
"command": "npx",
"args": [
"@browserbasehq/mcp"
],
"env": {
"BROWSERBASE_API_KEY": "<your Browserbase API key>",
"BROWSERBASE_PROJECT_ID": "<your Browserbase project ID>",
"GEMINI_API_KEY": "<your Gemini API key for the default model>"
}
}
}
}Config file · Project scope: .mcp.json at the repository root
The project file uses mcpServers. For user scope, use claude mcp add --scope user rather than assuming .mcp.json in your home directory is the user configuration. Check the client documentation for managed settings and storage details. Merge this example with existing settings; do not overwrite unrelated servers.
{
"mcpServers": {
"mcp-server-browserbase": {
"type": "stdio",
"command": "npx",
"args": [
"@browserbasehq/mcp"
],
"env": {
"BROWSERBASE_API_KEY": "<your Browserbase API key>",
"BROWSERBASE_PROJECT_ID": "<your Browserbase project ID>",
"GEMINI_API_KEY": "<your Gemini API key for the default model>"
}
}
}
}Examples are translated from the published configuration. Supply your own credentials and verify requirements in the upstream README. Remote connectors are shown only where a supported configuration format is known.
Tools
Source: repository README. Tool descriptions are a documented snapshot, not a live capability check.
| Tool | What it does |
|---|---|
| start | Create a Browserbase session or reuse an existing one. |
| end | Close the active Browserbase session. |
| navigate | Open a URL in the session browser. |
| act | Carry out an instructed action on the page. |
| observe | Identify page elements relevant to an interaction instruction. |
| extract | Retrieve page data, optionally guided by an instruction. |
Names, permissions, and defaults can change. Check the current upstream documentation before granting access.
Use cases
- Run a browser session from an assistant without managing a local browser.
- Navigate and act on live pages using instructions.
- Discover actionable page elements before choosing an operation.
- Extract page data through Stagehand-backed automation.
Strengths
- Provides the same six-tool interface in hosted and self-hosted variants.
- Offers direct HTTP access and a bridge for other clients.
- Self-hosted execution supports model and browser-session customization.
Limitations
- Self-hosted setup requires Browserbase credentials and model authentication.
- Custom models require a separate provider key.
- Verified Identity is restricted to Scale plan users.
Frequently asked questions
Independent listing, not an affiliation or security review. Grant the least access needed and verify the upstream instructions before running code.
