Ainative Browser MCP server
{
"mcpServers": {
"ainative-studio-ainative-browser": {
"command": "npx",
"args": [
"-y",
"@ainative/browser-mcp@latest"
],
"env": {
"AINATIVE_API_KEY": "<ainative_api_key>"
}
}
}
}{
"mcpServers": {
"ainative-studio-ainative-browser": {
"command": "npx",
"args": [
"-y",
"@ainative/browser-mcp@latest"
],
"env": {
"AINATIVE_API_KEY": "<ainative_api_key>"
}
}
}
}{
"mcpServers": {
"ainative-studio-ainative-browser": {
"command": "npx",
"args": [
"-y",
"@ainative/browser-mcp@latest"
],
"env": {
"AINATIVE_API_KEY": "<ainative_api_key>"
}
}
}
}{
"servers": {
"ainative-studio-ainative-browser": {
"command": "npx",
"args": [
"-y",
"@ainative/browser-mcp@latest"
],
"env": {
"AINATIVE_API_KEY": "${input:ainative_api_key}"
}
}
},
"inputs": [
{
"id": "ainative_api_key",
"type": "promptString",
"description": "Ainative api key",
"password": true
}
]
}{
"mcpServers": {
"ainative-studio-ainative-browser": {
"command": "npx",
"args": [
"-y",
"@ainative/browser-mcp@latest"
],
"env": {
"AINATIVE_API_KEY": "<ainative_api_key>"
}
}
}
}file .mcp.json (in your project root)file claude_desktop_config.jsonfile ~/.cursor/mcp.jsonfile .vscode/mcp.jsonfile ~/.codeium/windsurf/mcp_config.jsoninstall npx -y @ainative/browser-mcpversion 1.1.2
What it does
This MCP server enables AI agents to automate web browser interactions, extract structured data from pages, and validate content. It provides six tools for web automation tasks and can store extracted data in agent memory or database tables.
- Extract product information, pricing, or other structured data from websites and store it in a table.
- Perform multi-step browser actions like filling forms, clicking buttons, and navigating pages.
- Validate that specific content or page states exist as expected before proceeding with further actions.
Tools this server exposes
| Tool | Does |
|---|---|
| browser_act | Perform actions on a web page such as clicking, typing, or navigating. |
| browser_extract | Extract structured data from a page based on a goal description. |
| browser_validate | Validate that content or a specific state exists on a page. |
| browser_task | Run a multi-step browser automation task. |
| browser_extract_to_table | Extract data from a page and store it in a ZeroDB table. |
| browser_enrich_memory | Extract content from a page and store it in agent memory. |
Before you enable it
An MCP server runs with whatever access you give it. Start with a development account rather than production, and check which tools the server exposes before you let an agent call them unattended.