Datalab Document Converter

Convert PDF, EPUB, PPTX, DOCX, XLSX, HTML, and image files to Markdown using the Datalab cloud API.

Prerequisites

# Install Datalab Python SDK
uv pip install datalab-python-sdk

# Set API key (get from https://www.datalab.to)
export DATALAB_API_KEY="your_api_key_here"

Python SDK Usage

Basic Conversion

from datalab_sdk import DatalabClient

client = DatalabClient()  # Uses DATALAB_API_KEY env var

# Convert document to markdown
result = client.convert("document.pdf")
print(result.markdown)

# Save output
result = client.convert(
    "document.pdf",
    save_output="./output/document"
)
# Creates: output/document.md, output/document_meta.json, output/*.png

With Options

from datalab_sdk import DatalabClient, ConvertOptions

client = DatalabClient()

options = ConvertOptions(
    output_format="markdown",  # markdown, json, html, chunks
    force_ocr=False,           # Force OCR on all pages
    paginate=True,             # Add page separators
    use_llm=True,              # Use LLM for better accuracy
    disable_image_extraction=True,  # Plain text only
    page_range="0,5-10,20"     # Specific pages
)

result = client.convert("document.pdf", options=options)

Async Client (Better Performance)

import asyncio
from datalab_sdk import AsyncDatalabClient, ConvertOptions

async def convert_document():
    async with AsyncDatalabClient() as client:
        result = await client.convert(
            "document.pdf",
            options=ConvertOptions(output_format="markdown")
        )
        return result.markdown

markdown = asyncio.run(convert_document())
print(markdown)

OCR Only

from datalab_sdk import DatalabClient

client = DatalabClient()

# OCR a document
ocr_result = client.ocr("document.pdf")
print(ocr_result.pages)  # Get all text

REST API Usage

Submit Document for Conversion

import requests

url = "https://www.datalab.to/api/v1/marker"
headers = {"X-API-Key": "YOUR_API_KEY"}

with open("document.pdf", "rb") as f:
    files = {"file": ("document.pdf", f, "application/pdf")}
    data = {
        "output_format": (None, "markdown"),
        "force_ocr": (None, "false"),
        "use_llm": (None, "false"),
        "disable_image_extraction": (None, "true")
    }
    response = requests.post(url, headers=headers, files=files, data=data)

result = response.json()
print(f"Request ID: {result['request_id']}")
print(f"Check URL: {result['request_check_url']}")

Poll for Results

import requests
import time

check_url = result['request_check_url']
headers = {"X-API-Key": "YOUR_API_KEY"}

while True:
    response = requests.get(check_url, headers=headers)
    status = response.json()

    if status.get('status') == 'complete':
        print(status['markdown'])
        break
    elif status.get('status') == 'failed':
        print(f"Error: {status.get('error')}")
        break

    time.sleep(2)  # Poll every 2 seconds

Using curl

# Submit document
curl -X POST "https://www.datalab.to/api/v1/marker" \
  -H "X-API-Key: $DATALAB_API_KEY" \
  -F "file=@document.pdf" \
  -F "output_format=markdown" \
  -F "disable_image_extraction=true"

# Check status
curl "https://www.datalab.to/api/v1/marker/{request_id}" \
  -H "X-API-Key: $DATALAB_API_KEY"

API Options

| Parameter | Type | Description | | -------------------------- | ------- | ------------------------------------ | | output_format | string | markdown, json, html, chunks | | force_ocr | boolean | Force OCR on all pages | | paginate | boolean | Add page separators | | use_llm | boolean | Use LLM for better accuracy | | strip_existing_ocr | boolean | Remove existing OCR and re-process | | disable_image_extraction | boolean | Plain text only | | page_range | string | Specific pages, e.g., "0,5-10,20" | | max_pages | integer | Maximum pages to convert |

Batch Processing

import asyncio
from pathlib import Path
from datalab_sdk import AsyncDatalabClient, ConvertOptions

async def batch_convert(files: list[Path], output_dir: Path):
    output_dir.mkdir(parents=True, exist_ok=True)

    options = ConvertOptions(
        output_format="markdown",
        disable_image_extraction=True
    )

    async with AsyncDatalabClient() as client:
        tasks = [
            client.convert(
                file_path=f,
                options=options,
                save_output=output_dir / f.stem
            )
            for f in files
        ]
        results = await asyncio.gather(*tasks, return_exceptions=True)

    for f, result in zip(files, results):
        if isinstance(result, Exception):
            print(f"✗ {f.name}: {result}")
        elif result.success:
            print(f"✓ {f.name}: {result.page_count} pages")
        else:
            print(f"✗ {f.name}: {result.error}")

# Usage
files = list(Path("documents").glob("*.pdf"))
asyncio.run(batch_convert(files, Path("output")))

Error Handling

from datalab_sdk import (
    DatalabClient,
    DatalabAPIError,
    DatalabTimeoutError,
    DatalabFileError
)

client = DatalabClient()

try:
    result = client.convert("document.pdf", max_polls=60, poll_interval=2)

    if result.success:
        print(result.markdown)
    else:
        print(f"Conversion failed: {result.error}")

except DatalabAPIError as e:
    if e.status_code == 401:
        print("Authentication failed - check API key")
    elif e.status_code == 429:
        print("Rate limit exceeded - wait before retrying")
    else:
        print(f"API Error: {e}")

except DatalabTimeoutError:
    print("Operation timed out - try increasing max_polls")

except DatalabFileError as e:
    print(f"File error: {e}")

Datalab vs Marker CLI

| Feature | Datalab API | Marker CLI | | ------------ | ------------------ | ------------------- | | Processing | Cloud-based | Local | | GPU Required | No | Yes (recommended) | | Setup | API key only | Python + PyTorch | | Speed | Fast (cloud GPU) | Depends on hardware | | Privacy | Data sent to cloud | Local processing | | Cost | API credits | Free |

Instructions

1. Confirm the input file path exists

2. Check if $DATALAB_API_KEY environment variable is set

3. Use AskUserQuestion tool to ask user preferences:

   Question 1 - Processing Method:

   • Header: "Method"
   • Question: "使用哪种方式调用 Datalab API？"
   • Options:
     • "Python SDK (Recommended)": 使用 datalab-python-sdk，更简洁
     • "REST API": 使用 requests 直接调用 API
     • "curl": 使用命令行 curl

   Question 2 - Image Extraction:

   • Header: "Images"
   • Question: "是否需要提取文档中的图片？"
   • Options:
     • "No (Recommended)": 仅提取文本，生成纯 Markdown
     • "Yes": 提取图片并保存

4. Generate and run the appropriate code based on user's choice

5. Report the output file location and any extraction notes