Anakin MCP server
{
"mcpServers": {
"anakin": {
"command": "npx",
"args": [
"-y",
"@anakin-io/mcp@latest"
],
"env": {
"ANAKIN_API_KEY": "<anakin_api_key>"
}
}
}
}{
"mcpServers": {
"anakin": {
"command": "npx",
"args": [
"-y",
"@anakin-io/mcp@latest"
],
"env": {
"ANAKIN_API_KEY": "<anakin_api_key>"
}
}
}
}{
"mcpServers": {
"anakin": {
"command": "npx",
"args": [
"-y",
"@anakin-io/mcp@latest"
],
"env": {
"ANAKIN_API_KEY": "<anakin_api_key>"
}
}
}
}{
"servers": {
"anakin": {
"command": "npx",
"args": [
"-y",
"@anakin-io/mcp@latest"
],
"env": {
"ANAKIN_API_KEY": "${input:anakin_api_key}"
}
}
},
"inputs": [
{
"id": "anakin_api_key",
"type": "promptString",
"description": "Anakin api key",
"password": true
}
]
}{
"mcpServers": {
"anakin": {
"command": "npx",
"args": [
"-y",
"@anakin-io/mcp@latest"
],
"env": {
"ANAKIN_API_KEY": "<anakin_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 @anakin-io/mcpversion 0.2.1
What it does
MCP server that gives AI agents access to web scraping, searching, crawling, site monitoring, and browser automation through the Anakin API. Designed for developers building AI applications that need to interact with web data and automate tasks across hundreds of websites.
- Scrape and extract data from web pages as markdown or structured JSON without writing parsing code.
- Perform multi-source deep research or monitor web pages for changes on a schedule and receive alerts.
- Automate read and write actions on websites (like extracting product listings or submitting forms) through pre-built Wire actions.
Tools this server exposes
| Tool | Does |
|---|---|
| scrape | Fetch a single URL and return its content as markdown or AI-extracted JSON. |
| search | Perform AI web search with citations. |
| map | Discover all URLs on a domain. |
| crawl | Bulk-fetch and convert multiple pages on a site to markdown. |
| agentic_search | Multi-source deep research that runs for 1–5 minutes. |
| wire_discover | Find a Wire action for a task by natural-language description. |
| wire_catalog | Browse supported sites and their available actions with parameter schemas. |
| wire_read_action | Run a Wire read action to extract data from a website. |
| wire_write_action | Run a Wire write action to perform state-changing operations on a website. |
| wire_identities | List saved accounts and credentials for auth-required actions. |
| wire_login | Sign in to a credentials-required site and store the session. |
| wire_build | Request a new Wire action for a site not yet in the catalog. |
| monitor_create | Set up monitoring for a page, site, or Wire action and receive alerts on changes. |
| monitor_list | List all monitors or retrieve details for a specific monitor. |
| monitor_changes | Retrieve detected changes for a monitor. |
| monitor_control | Pause, resume, run immediately, or delete a monitor. |
| ai_visibility_search | Ask multiple AI answer engines (ChatGPT, Gemini, Google AI Overview) the same question and compare results. |
| ai_visibility_sources | List available AI answer engines for comparison queries. |
| session_list | List saved browser login sessions. |
| session_delete | Delete a saved browser login session. |
| browser_task | Execute a natural-language task in an AI-driven cloud browser for up to ~5 minutes. |
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.