> ## Documentation Index
> Fetch the complete documentation index at: https://gologin.com/docs/llms.txt
> Use this file to discover all available pages before exploring further.

# MCP Server

> Manage Gologin browser profiles through AI conversations via the Model Context Protocol

Gologin MCP Server lets you manage browser profiles through AI conversations using the [Model Context Protocol](https://modelcontextprotocol.io/). Connect it to Claude Desktop, Cursor, Claude Code, or any MCP-compatible client and control profiles with natural language.

***

## What you can do

* **Profile management** — create, update, delete, and list browser profiles
* **Proxy configuration** — set up proxies on profiles
* **Fingerprint control** — customize browser fingerprints and user agents
* **Folder organization** — create folders and organize profiles
* **Browser sessions** — start and stop browser sessions
* **Account info** — check subscription status and usage

***

## Install

### Requirements

* Node.js 18 or higher
* A Gologin API token
* An active Gologin account

### Get your API token

1. Log in at [gologin.com](https://gologin.com)
2. Open **API & MCP** in the dashboard
3. Open the **API** tab
4. Click **New Token** and copy it

### Claude Desktop

1. Open Claude Desktop → **Settings**
2. Click **Developer** in the sidebar, then **Edit Config**
3. Add to `claude_desktop_config.json`:

```json theme={null}
{
  "mcpServers": {
    "gologin-mcp": {
      "command": "npx",
      "args": ["gologin-mcp"],
      "env": {
        "API_TOKEN": "your-gologin-api-token-here"
      }
    }
  }
}
```

4. Restart Claude Desktop
5. Check **Settings → Connectors** for "gologin-mcp LOCAL"

### Claude Code

Add to your Claude Code MCP settings:

```json theme={null}
{
  "mcpServers": {
    "gologin-mcp": {
      "command": "npx",
      "args": ["gologin-mcp"],
      "env": {
        "API_TOKEN": "your-gologin-api-token-here"
      }
    }
  }
}
```

### Cursor

Use the same JSON configuration in Cursor's MCP settings file.

### Other MCP clients

Any client that supports the Model Context Protocol can use the same configuration structure.

***

## Usage examples

Once connected, talk to your AI assistant naturally:

**Account**

* "What's my Gologin account status?"
* "Check my subscription details"

**Creating profiles**

* "Create a new browser profile with a US proxy"
* "Set up 5 profiles for social media management"

**Managing profiles**

* "Show me all my profiles"
* "Update the proxy for profile ID 123 to use a UK proxy"
* "Delete the profile named 'test-profile'"

**Organization**

* "Create a folder called 'Social Media Accounts'"
* "Move profile XYZ to the Social Media folder"

**Browser sessions**

* "Start a browser session for my profile"
* "Stop all running browser sessions"

***

## Real-world scenarios

### Set up profiles for a new client

You're onboarding a new client who needs 10 browser profiles for different regional ad accounts:

> "Create 10 browser profiles named 'Client X - US', 'Client X - UK', 'Client X - DE' and so on. Set each one with a residential proxy from its respective country. Put them all in a folder called 'Client X'."

Claude creates the profiles, configures proxies, and organizes them — all through natural language.

### Audit your existing setup

> "Show me all profiles that don't have a proxy configured. Also, which profiles haven't been used in the last 30 days?"

Quick way to find misconfigured or abandoned profiles without clicking through the dashboard.

### Prepare profiles before automation

MCP Server handles the setup; CLI tools handle the execution:

1. **MCP:** "Create 5 profiles with US proxies for scraping, put them in a 'Scraping' folder"
2. **CLI:** `gologin-agent-browser open https://target.com --profile profile_1` to run automation
3. **MCP:** "Delete all profiles in the 'Scraping' folder" when done

### Team management

> "How many active sessions are running right now? What's my plan's session limit?"
> "Check my subscription — how many profiles do I have left?"

Useful for team leads who need to monitor usage without opening the dashboard.

***

## MCP Server vs CLI tools vs AI Skills

| Tool                                                  | Best for                                           | How it works                          |
| ----------------------------------------------------- | -------------------------------------------------- | ------------------------------------- |
| **MCP Server**                                        | Profile management, account operations             | Natural language through MCP protocol |
| [CLI Tools](/cli-tools/web-access)                    | Web scraping, browser automation, batch operations | Shell commands                        |
| [AI Skills](/api-reference/ai-integrations/ai-skills) | Plug-and-play agent capabilities in Claude Code    | Skill extensions wrapping CLI tools   |

MCP Server focuses on **profile and account management**. For web scraping and browser automation, use the [CLI tools](/cli-tools/web-access) or [AI Skills](/api-reference/ai-integrations/ai-skills).

***

## Related

* [AI Skills](/api-reference/ai-integrations/ai-skills) — plug-and-play web access skills for Claude Code
* [Web Access CLI](/cli-tools/web-access) — unified scraping + browser CLI
* [Agent Browser CLI](/cli-tools/agent-browser) — cloud browser CLI for AI agents

<Card title="npm package" icon="npm" href="https://www.npmjs.com/package/gologin-mcp">
  gologin-mcp
</Card>

<Card title="Source code" icon="github" href="https://github.com/gologinapp/gologin-mcp">
  gologinapp/gologin-mcp
</Card>