# Installation with pip and pipx
This guide covers traditional Python package installation methods (pip and pipx) for Python Alfresco MCP Server. For modern UV/UVX installation (recommended), see the [main README](../README.md).
## ๐ ๏ธ Installation Options
### Option 1: pipx (Recommended for Traditional Methods)
pipx automatically creates isolated environments for each tool while making commands globally available - eliminates dependency conflicts while providing system-wide access.
```bash
# First install pipx if you don't have it (one-time setup)
pip install pipx
# Install python-alfresco-mcp-server in isolated environment
pipx install python-alfresco-mcp-server
# Test that installation worked
python-alfresco-mcp-server --help
```
**Why pipx?** pipx automatically creates isolated environments for each tool while making commands globally available - eliminates dependency conflicts while providing system-wide access.
### Option 2: pip (Traditional Package Manager)
```bash
# Recommended: Create virtual environment first
python -m venv venv
source venv/bin/activate # Linux/macOS
# venv\Scripts\activate # Windows
# Install python-alfresco-mcp-server
pip install python-alfresco-mcp-server
# Test that installation worked
python-alfresco-mcp-server --help
```
## ๐ Usage
### MCP Server Startup
**With pipx (Global installation - no venv needed):**
```bash
# Run MCP server with STDIO transport (default)
python-alfresco-mcp-server
# HTTP transport for web services (matches MCP Inspector)
python-alfresco-mcp-server --transport http --host 127.0.0.1 --port 8003
# SSE transport for real-time streaming
python-alfresco-mcp-server --transport sse --host 127.0.0.1 --port 8001
```
**With pip (Activate venv first if installed in one):**
```bash
# Activate virtual environment first (if used during installation)
source venv/bin/activate # Linux/macOS
# venv\Scripts\activate # Windows
# Run MCP server with STDIO transport (default)
python-alfresco-mcp-server
# HTTP transport for web services (matches MCP Inspector)
python-alfresco-mcp-server --transport http --host 127.0.0.1 --port 8003
# SSE transport for real-time streaming
python-alfresco-mcp-server --transport sse --host 127.0.0.1 --port 8001
```
## ๐ง Claude Desktop Configuration
The Claude Desktop configuration differs based on how you installed the MCP server:
### pipx Installation (Global Tool)
```json
{
"command": "python-alfresco-mcp-server",
"args": ["--transport", "stdio"]
}
```
- Uses the **global command name** directly (no path needed)
- pipx makes tools globally available in your PATH
- Simplest configuration
**Sample Config Files:**
- Windows: [`claude-desktop-config-pipx-windows.json`](../claude-desktop-config-pipx-windows.json)
- macOS: [`claude-desktop-config-pipx-macos.json`](../claude-desktop-config-pipx-macos.json)
### pip Installation (Manual venv)
```json
{
"command": "C:\\path\\to\\venv\\Scripts\\python-alfresco-mcp-server.exe",
"args": ["--transport", "stdio"]
}
```
- Uses **direct path to executable** in your virtual environment
- Path points to `Scripts/` directory in your venv (Windows) or `bin/` (Linux/macOS)
- Replace `C:\\path\\to\\venv` with your actual venv location
## ๐ MCP Inspector Configuration
For development and testing with MCP Inspector:
### pipx Installation
Use the sample config files:
- **stdio transport**: [`mcp-inspector-stdio-pipx-config.json`](../mcp-inspector-stdio-pipx-config.json)
- **http transport**: [`mcp-inspector-http-pipx-config.json`](../mcp-inspector-http-pipx-config.json)
```bash
# Start with stdio transport
npx @modelcontextprotocol/inspector --config mcp-inspector-stdio-pipx-config.json --server python-alfresco-mcp-server
# Start with http transport
npx @modelcontextprotocol/inspector --config mcp-inspector-http-pipx-config.json --server python-alfresco-mcp-server
```
### pip Installation
Copy one of the pipx sample files above and modify the `"command"` field to point to your venv executable:
- Change `"python-alfresco-mcp-server"` to `"C:\\path\\to\\venv\\Scripts\\python-alfresco-mcp-server.exe"` (Windows)
- Or `"/path/to/venv/bin/python-alfresco-mcp-server"` (Linux/macOS)