# Quick Start: Build and Transfer Docker Image
**Use Case**: Build Docker image on Linux, transfer to Windows workstation for MCP client deployment.
## On Linux Build Machine
```bash
cd /path/to/mcpragcrawl4ai/deployments/client
# Build and export Docker image
./build-image.sh
# This creates: crawl4ai-rag-client.tar (~200-300MB)
```
## Transfer to Windows
**Copy these files to your Windows workstation:**
1. `crawl4ai-rag-client.tar` (Docker image)
2. `.env_template.txt` (config template)
3. `docker-compose.yml` (container setup)
4. `deploy-on-windows.ps1` (deployment script)
**Methods**: USB drive, SCP, network share, cloud storage
## On Windows Workstation
**Open PowerShell in the directory with transferred files:**
```powershell
# Run deployment script
.\deploy-on-windows.ps1
# Follow prompts to:
# - Load Docker image
# - Configure .env
# - Start container
```
**Configure Claude Desktop:**
Edit: `%APPDATA%\Claude\claude_desktop_config.json`
```json
{
"mcpServers": {
"crawl4ai-rag": {
"command": "docker",
"args": ["exec", "-i", "crawl4ai-rag-client", "python", "-m", "core.rag_processor"]
}
}
}
```
**Restart Claude Desktop** - Tools will appear!
---
**See [IMAGE-TRANSFER.md](IMAGE-TRANSFER.md) for detailed instructions and troubleshooting.**
