# Directory Structure
```
├── .env
├── .gitignore
├── .python-version
├── Dockerfile
├── images
│ ├── input.png
│ └── output.png
├── LICENSE
├── pyproject.toml
├── README_CN.md
├── README.md
├── smithery.yaml
├── src
│ └── pdf2md
│ ├── __init__.py
│ ├── __main__.py
│ └── server.py
└── uv.lock
```
# Files
--------------------------------------------------------------------------------
/.python-version:
--------------------------------------------------------------------------------
```
3.10
```
--------------------------------------------------------------------------------
/.env:
--------------------------------------------------------------------------------
```
MINERU_API_BASE = "https://mineru.net/api/v4/extract/task"
MINERU_API_KEY = "Your API_KEY"
MINERU_FILE_URLS_API = "https://mineru.net/api/v4/file-urls/batch"
MINERU_BATCH_API = "https://mineru.net/api/v4/extract/task/batch"
MINERU_BATCH_RESULTS_API = "https://mineru.net/api/v4/extract-results/batch"
```
--------------------------------------------------------------------------------
/README.md:
--------------------------------------------------------------------------------
```markdown
# MCP-PDF2MD
[](https://smithery.ai/server/@FutureUnreal/mcp-pdf2md)
[English](#pdf2md-service) | [中文](README_CN.md)
# MCP-PDF2MD Service
An MCP-based high-performance PDF to Markdown conversion service powered by MinerU API, supporting batch processing for local files and URL links with structured output.
## Key Features
- Format Conversion: Convert PDF files to structured Markdown format.
- Multi-source Support: Process both local PDF files and URL links.
- Intelligent Processing: Automatically select the best processing method.
- Batch Processing: Support multi-file batch conversion for efficient handling of large volumes of PDF files.
- MCP Integration: Seamless integration with LLM clients like Claude Desktop.
- Structure Preservation: Maintain the original document structure, including headings, paragraphs, lists, etc.
- Smart Layout: Output text in human-readable order, suitable for single-column, multi-column, and complex layouts.
- Formula Conversion: Automatically recognize and convert formulas in the document to LaTeX format.
- Table Extraction: Automatically recognize and convert tables in the document to structured format.
- Cleanup Optimization: Remove headers, footers, footnotes, page numbers, etc., to ensure semantic coherence.
- High-Quality Extraction: High-quality extraction of text, images, and layout information from PDF documents.
## System Requirements
- Software: Python 3.10+
## Quick Start
1. Clone the repository and enter the directory:
```bash
git clone https://github.com/FutureUnreal/mcp-pdf2md.git
cd mcp-pdf2md
```
2. Create a virtual environment and install dependencies:
**Linux/macOS**:
```bash
uv venv
source .venv/bin/activate
uv pip install -e .
```
**Windows**:
```bash
uv venv
.venv\Scripts\activate
uv pip install -e .
```
3. Configure environment variables:
Create a `.env` file in the project root directory and set the following environment variables:
```
MINERU_API_BASE=https://mineru.net/api/v4/extract/task
MINERU_BATCH_API=https://mineru.net/api/v4/extract/task/batch
MINERU_BATCH_RESULTS_API=https://mineru.net/api/v4/extract-results/batch
MINERU_API_KEY=your_api_key_here
```
4. Start the service:
```bash
uv run pdf2md
```
## Command Line Arguments
The server supports the following command line arguments:
## Claude Desktop Configuration
Add the following configuration in Claude Desktop:
**Windows**:
```json
{
"mcpServers": {
"pdf2md": {
"command": "uv",
"args": [
"--directory",
"C:\\path\\to\\mcp-pdf2md",
"run",
"pdf2md",
"--output-dir",
"C:\\path\\to\\output"
],
"env": {
"MINERU_API_KEY": "your_api_key_here"
}
}
}
}
```
**Linux/macOS**:
```json
{
"mcpServers": {
"pdf2md": {
"command": "uv",
"args": [
"--directory",
"/path/to/mcp-pdf2md",
"run",
"pdf2md",
"--output-dir",
"/path/to/output"
],
"env": {
"MINERU_API_KEY": "your_api_key_here"
}
}
}
}
```
**Note about API Key Configuration:**
You can set the API key in two ways:
1. In the `.env` file within the project directory (recommended for development)
2. In the Claude Desktop configuration as shown above (recommended for regular use)
If you set the API key in both places, the one in the Claude Desktop configuration will take precedence.
## MCP Tools
The server provides the following MCP tools:
- **convert_pdf_url**: Convert PDF URL to Markdown
- **convert_pdf_file**: Convert local PDF file to Markdown
## Getting MinerU API Key
This project relies on the MinerU API for PDF content extraction. To obtain an API key:
1. Visit [MinerU official website](https://mineru.net/) and register for an account
2. After logging in, apply for API testing qualification at [this link](https://mineru.net/apiManage/docs?openApplyModal=true)
3. Once your application is approved, you can access the [API Management](https://mineru.net/apiManage/token) page
4. Generate your API key following the instructions provided
5. Copy the generated API key
6. Use this string as the value for `MINERU_API_KEY`
Note that access to the MinerU API is currently in testing phase and requires approval from the MinerU team. The approval process may take some time, so plan accordingly.
## Demo
### Input PDF

### Output Markdown

## License
MIT License - see the LICENSE file for details.
## Credits
This project is based on the API from [MinerU](https://github.com/opendatalab/MinerU/tree/master).
```
--------------------------------------------------------------------------------
/src/pdf2md/__main__.py:
--------------------------------------------------------------------------------
```python
"""
PDF to Markdown Conversion Service Startup Script
"""
import asyncio
from . import main
if __name__ == "__main__":
main()
```
--------------------------------------------------------------------------------
/pyproject.toml:
--------------------------------------------------------------------------------
```toml
[project]
name = "pdf2md"
version = "0.1.0"
description = "PDF to Markdown MCP服务器"
readme = "README.md"
requires-python = ">=3.10"
dependencies = [
"httpx>=0.28.1",
"mcp[cli]>=1.4.1",
"python-dotenv>=1.0.0",
"asyncio>=3.4.3",
"pathlib>=1.0.1",
"typer>=0.9.0",
]
[project.optional-dependencies]
dev = [
"pytest>=7.0.0",
]
[project.scripts]
pdf2md = "pdf2md:main"
[build-system]
requires = ["setuptools>=61.0"]
build-backend = "setuptools.build_meta"
[tool.hatch.build.targets.wheel]
packages = ["src/pdf2md"]
```
--------------------------------------------------------------------------------
/Dockerfile:
--------------------------------------------------------------------------------
```dockerfile
# Generated by https://smithery.ai. See: https://smithery.ai/docs/config#dockerfile
FROM python:3.10-slim
# Install system dependencies if needed
RUN apt-get update && apt-get install -y --no-install-recommends \
build-essential \
&& rm -rf /var/lib/apt/lists/*
WORKDIR /app
# Copy the entire project to /app
COPY . /app
# Upgrade pip and install the package in editable mode
RUN pip install --upgrade pip
RUN pip install -e .
# Expose any necessary ports if applicable (none required for stdio based MCP)
# Default command to run the MCP server using python module entrypoint
CMD ["python", "-m", "./src/pdf2md"]
```
--------------------------------------------------------------------------------
/src/pdf2md/__init__.py:
--------------------------------------------------------------------------------
```python
from .server import mcp
def main():
"""PDF to Markdown Conversion Service - Provides MCP service for converting PDF files to Markdown"""
import os
import argparse
# Parse command line arguments
parser = argparse.ArgumentParser(description="PDF to Markdown Conversion Service")
parser.add_argument("--output-dir", default="./downloads", help="Specify output directory path, default is ./downloads")
args = parser.parse_args()
# Set output directory
from .server import set_output_dir
set_output_dir(args.output_dir)
# Check API key
from .server import MINERU_API_KEY, logger
if not MINERU_API_KEY:
logger.warning("Warning: API key not set, please set the MINERU_API_KEY environment variable")
# Run MCP server
mcp.run()
__all__ = ['main']
```
--------------------------------------------------------------------------------
/smithery.yaml:
--------------------------------------------------------------------------------
```yaml
# Smithery configuration file: https://smithery.ai/docs/config#smitheryyaml
startCommand:
type: stdio
configSchema:
# JSON Schema defining the configuration options for the MCP.
type: object
required:
- mineruApiKey
properties:
mineruApiKey:
type: string
description: API key for MinerU. Must be provided as a Bearer token without the
'Bearer ' prefix.
outputDir:
type: string
default: ./downloads
description: Directory where the converted markdown files will be stored.
commandFunction:
# A JS function that produces the CLI command based on the given config to start the MCP on stdio.
|-
(config) => ({
command: 'python',
args: ['-m', './src/pdf2md', '--output-dir', config.outputDir],
env: { MINERU_API_KEY: config.mineruApiKey }
})
exampleConfig:
mineruApiKey: your_mineru_api_key_here
outputDir: ./downloads
```
--------------------------------------------------------------------------------
/src/pdf2md/server.py:
--------------------------------------------------------------------------------
```python
import os
import json
import time
import asyncio
import httpx
import re
import logging
from pathlib import Path
from dotenv import load_dotenv
from typing import Optional, List, Dict, Any
from mcp.server.fastmcp import FastMCP
# Set up logging - disable all log output
logging.basicConfig(
level=logging.CRITICAL, # Only record critical errors
format='%(asctime)s - %(name)s - %(levelname)s - %(message)s',
handlers=[] # Remove all handlers, no log output
)
logger = logging.getLogger("pdf2md")
logger.disabled = True # Completely disable logging
# Load environment variables
load_dotenv()
# API configuration
MINERU_API_BASE = os.environ.get("MINERU_API_BASE", "https://mineru.net/api/v4/extract/task")
MINERU_API_KEY = os.environ.get("MINERU_API_KEY", "")
MINERU_BATCH_API = os.environ.get("MINERU_BATCH_API", "https://mineru.net/api/v4/extract/task/batch")
MINERU_BATCH_RESULTS_API = os.environ.get("MINERU_BATCH_RESULTS_API", "https://mineru.net/api/v4/extract-results/batch")
MINERU_FILE_URLS_API = os.environ.get("MINERU_FILE_URLS_API", "https://mineru.net/api/v4/file-urls/batch")
# Global variables
OUTPUT_DIR = "./downloads"
# API authentication headers
HEADERS = {
"Authorization": MINERU_API_KEY if MINERU_API_KEY.startswith("Bearer ") else f"Bearer {MINERU_API_KEY}",
"Content-Type": "application/json"
}
def set_output_dir(output_dir: str):
"""Set the output directory path"""
global OUTPUT_DIR
# Normalize path to handle Unicode characters properly
OUTPUT_DIR = os.path.normpath(output_dir)
def print_task_status(extract_results):
"""
Print task status and check if all tasks are completed
Args:
extract_results: List of task results
Returns:
tuple: (all tasks completed, any task completed)
"""
all_done = True
any_done = False
for i, result in enumerate(extract_results):
current_status = result.get("state", "")
file_name = result.get("file_name", "")
status_icon = "✅" if current_status == "done" else "⏳"
if current_status == "done":
any_done = True
else:
all_done = False
return all_done, any_done
async def check_task_status(client, batch_id, max_retries=60, sleep_seconds=5):
"""
Check batch task status
Args:
client: HTTP client
batch_id: Batch ID
max_retries: Maximum number of retries
sleep_seconds: Seconds between retries
Returns:
dict: Dictionary containing task status information, or error message if failed
"""
retry_count = 0
while retry_count < max_retries:
retry_count += 1
try:
status_response = await client.get(
f"{MINERU_BATCH_RESULTS_API}/{batch_id}",
headers=HEADERS,
timeout=60.0
)
if status_response.status_code != 200:
retry_count += 1
if retry_count < max_retries:
await asyncio.sleep(sleep_seconds)
continue
try:
status_data = status_response.json()
except Exception as e:
retry_count += 1
if retry_count < max_retries:
await asyncio.sleep(sleep_seconds)
continue
task_data = status_data.get("data", {})
extract_results = task_data.get("extract_result", [])
all_done, any_done = print_task_status(extract_results)
if all_done:
return {
"success": True,
"extract_results": extract_results,
"task_data": task_data,
"status_data": status_data
}
await asyncio.sleep(sleep_seconds)
except Exception as e:
retry_count += 1
if retry_count < max_retries:
await asyncio.sleep(sleep_seconds)
return {
"success": False,
"error": "Polling timeout, unable to get final results"
}
async def download_batch_results(client, extract_results):
"""
Download batch task results
Args:
client: HTTP client
extract_results: List of task results
Returns:
list: List of downloaded file information
"""
downloaded_files = []
for i, result in enumerate(extract_results):
if result.get("state") == "done":
try:
file_name = result.get("file_name", f"file_{i+1}")
zip_url = result.get("full_zip_url", "")
if not zip_url:
continue
downloaded_file = await download_zip_file(client, zip_url, file_name)
if downloaded_file:
downloaded_files.append(downloaded_file)
except Exception as e:
pass
return downloaded_files
async def download_zip_file(client, zip_url, file_name, prefix="md", max_retries=3):
"""
Download and save ZIP file, then automatically unzip
Args:
client: HTTP client
zip_url: ZIP file URL
file_name: File name
prefix: File prefix
max_retries: Maximum number of retries
Returns:
dict: Dictionary containing file name and unzip directory, or None if failed
"""
retry_count = 0
while retry_count < max_retries:
try:
zip_response = await client.get(zip_url, follow_redirects=True, timeout=120.0)
if zip_response.status_code == 200:
current_date = time.strftime("%Y%m%d")
base_name = os.path.splitext(file_name)[0]
# Only remove control characters and chars that are invalid in filenames
safe_name = re.sub(r'[<>:"/\\|?*\x00-\x1f]', '', base_name).strip()
# Replace spaces with underscores
safe_name = re.sub(r'\s+', '_', safe_name)
if safe_name.isdigit() or re.match(r'^\d+\.\d+$', safe_name):
safe_name = f"paper_{safe_name}"
zip_filename = f"{prefix}_{safe_name}_{current_date}.zip"
download_dir = Path(OUTPUT_DIR)
if not download_dir.exists():
try:
download_dir.mkdir(parents=True, exist_ok=True)
except Exception as e:
print(f"Error creating directory: {e}")
return None
save_path = download_dir / zip_filename
with open(save_path, "wb") as f:
f.write(zip_response.content)
extract_dir = download_dir / safe_name
if not extract_dir.exists():
extract_dir.mkdir(parents=True)
import zipfile
try:
with zipfile.ZipFile(save_path, 'r') as zip_ref:
zip_ref.extractall(extract_dir)
os.remove(save_path)
return {
"file_name": file_name,
"extract_dir": str(extract_dir)
}
except Exception as e:
pass
else:
retry_count += 1
if retry_count < max_retries:
await asyncio.sleep(2)
continue
except Exception as e:
retry_count += 1
if retry_count < max_retries:
await asyncio.sleep(2)
continue
return None
def parse_url_string(url_string):
"""
Parse URL string separated by spaces, commas, or newlines
Args:
url_string: URL string
Returns:
list: List of URLs
"""
if isinstance(url_string, str):
if (url_string.startswith('"') and url_string.endswith('"')) or \
(url_string.startswith("'") and url_string.endswith("'")):
url_string = url_string[1:-1]
urls = []
for part in url_string.split():
if ',' in part:
urls.extend(part.split(','))
elif '\n' in part:
urls.extend(part.split('\n'))
else:
urls.append(part)
cleaned_urls = []
for url in urls:
if (url.startswith('"') and url.endswith('"')) or \
(url.startswith("'") and url.endswith("'")):
cleaned_urls.append(url[1:-1])
else:
cleaned_urls.append(url)
return cleaned_urls
def parse_path_string(path_string):
"""
Parse file path string separated by spaces, commas, or newlines
Args:
path_string: File path string
Returns:
list: List of file paths
"""
if isinstance(path_string, str):
if (path_string.startswith('"') and path_string.endswith('"')) or \
(path_string.startswith("'") and path_string.endswith("'")):
path_string = path_string[1:-1]
paths = []
for part in path_string.split():
if ',' in part:
paths.extend(part.split(','))
elif '\n' in part:
paths.extend(part.split('\n'))
else:
paths.append(part)
cleaned_paths = []
for path in paths:
if (path.startswith('"') and path.endswith('"')) or \
(path.startswith("'") and path.endswith("'")):
cleaned_paths.append(path[1:-1])
else:
cleaned_paths.append(path)
return cleaned_paths
# Create MCP server
mcp = FastMCP("PDF to Markdown Conversion Service")
@mcp.tool()
async def convert_pdf_url(url: str, enable_ocr: bool = True) -> Dict[str, Any]:
"""
Convert PDF URL to Markdown, supports single URL or URL list
Args:
url: PDF file URL or URL list, can be separated by spaces, commas, or newlines
enable_ocr: Whether to enable OCR (default: True)
Returns:
dict: Conversion result information
"""
if not MINERU_API_KEY:
return {"success": False, "error": "Missing API key, please set environment variable MINERU_API_KEY"}
if isinstance(url, str):
urls = parse_url_string(url)
else:
urls = [url]
async with httpx.AsyncClient(timeout=300.0) as client:
try:
files = []
for i, url_item in enumerate(urls):
files.append({
"url": url_item,
"is_ocr": enable_ocr,
"data_id": f"url_convert_{i+1}_{int(time.time())}"
})
batch_data = {
"enable_formula": True,
"language": "auto",
"layout_model": "doclayout_yolo",
"enable_table": True,
"files": files
}
response = await client.post(
MINERU_BATCH_API,
headers=HEADERS,
json=batch_data,
timeout=300.0
)
if response.status_code != 200:
return {"success": False, "error": f"Request failed: {response.status_code}"}
try:
status_data = response.json()
if status_data.get("code") != 0 and status_data.get("code") != 200:
error_msg = status_data.get("msg", "Unknown error")
return {"success": False, "error": f"API returned error: {error_msg}"}
batch_id = status_data.get("data", {}).get("batch_id", "")
if not batch_id:
return {"success": False, "error": "Failed to get batch ID"}
task_status = await check_task_status(client, batch_id)
if not task_status.get("success"):
return task_status
downloaded_files = await download_batch_results(client, task_status.get("extract_results", []))
return {
"success": True,
"downloaded_files": downloaded_files,
"batch_id": batch_id,
"total_urls": len(urls),
"processed_urls": len(downloaded_files)
}
except json.JSONDecodeError as e:
return {"success": False, "error": f"Failed to parse JSON: {e}"}
except Exception as e:
return {"success": False, "error": str(e)}
@mcp.tool()
async def convert_pdf_file(file_path: str, enable_ocr: bool = True) -> Dict[str, Any]:
"""
Convert local PDF file to Markdown, supports single file or file list
Args:
file_path: PDF file local path or path list, can be separated by spaces, commas, or newlines
enable_ocr: Whether to enable OCR (default: True)
Returns:
dict: Conversion result information
"""
if not MINERU_API_KEY:
return {"success": False, "error": "Missing API key, please set environment variable MINERU_API_KEY"}
if isinstance(file_path, str):
file_paths = parse_path_string(file_path)
else:
file_paths = [file_path]
for path in file_paths:
if not os.path.exists(path):
return {"success": False, "error": f"File does not exist: {path}"}
else:
if not path.lower().endswith('.pdf'):
return {"success": False, "error": f"File is not in PDF format: {path}"}
async with httpx.AsyncClient(timeout=300.0) as client:
try:
file_names = [os.path.basename(path) for path in file_paths]
files_data = []
for i, name in enumerate(file_names):
files_data.append({
"name": name,
"is_ocr": enable_ocr,
"data_id": f"file_convert_{i+1}_{int(time.time())}"
})
file_url_data = {
"enable_formula": True,
"language": "auto",
"layout_model": "doclayout_yolo",
"enable_table": True,
"files": files_data
}
file_url_response = await client.post(
MINERU_FILE_URLS_API,
headers=HEADERS,
json=file_url_data,
timeout=60.0
)
if file_url_response.status_code != 200:
return {"success": False, "error": f"Failed to get upload link: {file_url_response.status_code}"}
file_url_result = file_url_response.json()
if file_url_result.get("code") != 0 and file_url_result.get("code") != 200:
error_msg = file_url_result.get("msg", "Unknown error")
return {"success": False, "error": f"Failed to get upload link: {error_msg}"}
batch_id = file_url_result.get("data", {}).get("batch_id", "")
file_urls = file_url_result.get("data", {}).get("file_urls", [])
if not batch_id or not file_urls or len(file_urls) != len(file_paths):
return {"success": False, "error": "Failed to get upload link or batch ID"}
upload_results = []
for i, (file_path, upload_url) in enumerate(zip(file_paths, file_urls)):
try:
with open(file_path, 'rb') as f:
file_content = f.read()
upload_response = await client.put(
upload_url,
content=file_content,
headers={},
timeout=300.0
)
if upload_response.status_code != 200:
upload_results.append({"file": file_names[i], "success": False})
else:
upload_results.append({"file": file_names[i], "success": True})
except Exception as e:
upload_results.append({"file": file_names[i], "success": False, "error": str(e)})
if not any(result["success"] for result in upload_results):
return {"success": False, "error": "All files failed to upload", "upload_results": upload_results}
task_status = await check_task_status(client, batch_id)
if not task_status.get("success"):
return task_status
downloaded_files = await download_batch_results(client, task_status.get("extract_results", []))
return {
"success": True,
"downloaded_files": downloaded_files,
"batch_id": batch_id,
"upload_results": upload_results,
"total_files": len(file_paths),
"processed_files": len(downloaded_files)
}
except Exception as e:
return {"success": False, "error": str(e)}
@mcp.prompt()
def default_prompt() -> str:
"""Create default tool usage prompt"""
return """
PDF to Markdown Conversion Service provides two tools, each with different functions:
- convert_pdf_url: Specifically designed for handling single or multiple URL links
- convert_pdf_file: Specifically designed for handling single or multiple local file paths
Please choose the appropriate tool based on the input type:
- If it's a single or multiple URL, use convert_pdf_url
- If it's a single or multiple local file, use convert_pdf_file
- If it's a mix of URL and local file, please call the above two tools separately to handle the corresponding input
The converted Markdown files will be saved in the specified output directory, and the temporary downloaded ZIP files will be automatically deleted after unzipping to save space.
"""
@mcp.prompt()
def pdf_prompt(path: str) -> str:
"""Create PDF processing prompt"""
return f"""
Please convert the following PDF to Markdown format:
{path}
Please choose the appropriate tool based on the input type:
- If it's a single or multiple URL, use convert_pdf_url
- If it's a single or multiple local file, use convert_pdf_file
The converted Markdown files will be saved in the specified output directory, and the temporary downloaded ZIP files will be automatically deleted after unzipping to save space.
"""
@mcp.resource("status://api")
def get_api_status() -> str:
"""Get API status information"""
if not MINERU_API_KEY:
return "API status: Not configured (missing API key)"
return f"API status: Configured\nAPI base URL: {MINERU_API_BASE}\nAPI key: {MINERU_API_KEY[:10]}..."
@mcp.resource("help://usage")
def get_usage_help() -> str:
"""Get tool usage help information"""
return """
# PDF to Markdown Conversion Service
## Available tools:
1. **convert_pdf_url** - Convert PDF URL to Markdown, supports single or multiple URLs
- Parameters:
- url: PDF file URL or URL list, can be separated by spaces, commas, or newlines
- enable_ocr: Whether to enable OCR (default: True)
2. **convert_pdf_file** - Convert local PDF file to Markdown, supports single or multiple file paths
- Parameters:
- file_path: PDF file local path or path list, can be separated by spaces, commas, or newlines
- enable_ocr: Whether to enable OCR (default: True)
## Tool functions:
- **convert_pdf_url**: Specifically designed for handling URL links, suitable for single or multiple URL inputs
- **convert_pdf_file**: Specifically designed for handling local files, suitable for single or multiple file path inputs
## Mixed input handling:
When handling both URL and local file inputs, please call the above two tools separately to handle the corresponding input parts.
## Usage examples:
```python
# Convert single URL
result = await convert_pdf_url("https://example.com/document.pdf")
# Convert multiple URLs (batch processing)
result = await convert_pdf_url('''
https://example.com/document1.pdf
https://example.com/document2.pdf
https://example.com/document3.pdf
''')
# Convert multiple URLs with comma separation
result = await convert_pdf_url("https://example.com/doc1.pdf, https://example.com/doc2.pdf")
# Convert single local file
result = await convert_pdf_file("C:/Documents/document.pdf")
# Convert multiple local files (batch processing)
result = await convert_pdf_file('''
C:/Documents/document1.pdf
C:/Documents/document2.pdf
C:/Documents/document3.pdf
''')
# Mixed input handling (URLs and local files)
url_result = await convert_pdf_url('''
https://example.com/doc1.pdf
https://example.com/doc2.pdf
''')
file_result = await convert_pdf_file('''
C:/Documents/doc1.pdf
C:/Documents/doc2.pdf
''')
```
## Conversion results:
Successful conversion returns a dictionary containing conversion result information, and the converted Markdown files will be saved in the specified output directory, with temporary downloaded ZIP files automatically deleted after unzipping to save space.
"""
if __name__ == "__main__":
if not MINERU_API_KEY:
print("Warning: API key not set, please set environment variable MINERU_API_KEY")
mcp.run()
```