# LM Studio MCP Configuration
This configuration enables OCR PDF MCP server integration with LM Studio.
## Prerequisites
1. Install Python dependencies:
```bash
pip install -r requirements.txt
```
2. Start the MCP server:
```bash
python run.py
```
## LM Studio Configuration
### Method 1: Direct MCP Integration
1. Open LM Studio
2. Go to **Settings** → **Extensions** → **MCP Servers**
3. Add new MCP server with these settings:
**Server Name**: `ocr-pdf`
**Command**: `python`
**Arguments**:
```json
["D:/AI/MCP/python/ocr_pdf_mcp/mcp_server_stdio_fixed.py"]
```
**Environment Variables**:
```
PYTHONPATH=D:\AI\MCP\python\ocr_pdf_mcp
```
### Method 2: HTTP Server Configuration
1. Start the HTTP server:
```bash
python run.py
```
2. In LM Studio, configure HTTP connection:
- **URL**: `http://localhost:8000`
- **Protocol**: `HTTP`
- **Content-Type**: `application/json`
### Method 3: JSON-RPC Configuration
Configure LM Studio to use JSON-RPC endpoint:
**Endpoint**: `http://localhost:8000/jsonrpc`
**Method**: `POST`
**Headers**:
```json
{
"Content-Type": "application/json"
}
```
## Available Tools
- **process_pdf_smart**: Intelligent PDF processing with automatic OCR detection
- **extract_pdf_text**: Direct text extraction from PDFs
- **perform_ocr**: OCR on image files (PNG, JPG, etc.)
- **analyze_pdf_structure**: Analyze PDF document structure and metadata
- **batch_process_pdfs**: Process multiple PDFs in batch mode
## Usage Examples
### Smart PDF Processing
```json
{
"jsonrpc": "2.0",
"method": "tools/call",
"params": {
"name": "process_pdf_smart",
"arguments": {
"pdf_path": "C:\\path\\to\\document.pdf",
"language": "eng"
}
},
"id": 1
}
```
### OCR Image
```json
{
"jsonrpc": "2.0",
"method": "tools/call",
"params": {
"name": "perform_ocr",
"arguments": {
"image_path": "C:\\path\\to\\image.png",
"language": "eng"
}
},
"id": 2
}
```
## Troubleshooting
1. **Connection Issues**: Ensure server is running on correct port
2. **Permission Errors**: Check file path permissions
3. **Tesseract Issues**: Install Tesseract OCR and add to PATH
## Repository
GitHub: https://github.com/irev/mcp-readpdfx
