# RAG Markdown Cleaner (`cooldev/rag-markdown-cleaner`) Actor

Transform web pages into RAG-ready Markdown with smart chunking, metadata, code detection & quality scoring. Production-tested deduplication. Fully open-source (Apache 2.0)—review code, contribute, or self-host. Turn messy HTML into embedding-ready knowledge instantly.

- **URL**: https://apify.com/cooldev/rag-markdown-cleaner.md
- **Developed by:** [Mohamed khalil Zouitni](https://apify.com/cooldev) (community)
- **Categories:** AI, Developer tools, Open source
- **Features:** Limited permissions
- **Stats:** 6 total users, 2 monthly users, 100.0% runs succeeded, 1 bookmarks
- **User rating**: No ratings yet

## Pricing

from $2.00 / 1,000 markdown results

This Actor is paid per event. You are not charged for the Apify platform usage, but only a fixed price for specific events.
Since this Actor supports Apify Store discounts, the price gets lower the higher subscription plan you have.

Learn more: https://docs.apify.com/actors/running/actors-in-store.md#pay-per-event

## What's an Apify Actor?

An Actor is a serverless cloud program that runs on the Apify platform. It has two run modes.
In Batch mode, an Actor accepts a well-defined JSON input, performs an action which can take anything from a few seconds to a few hours,
and optionally produces a well-defined JSON output, datasets with results, or files in key-value store.
In Standby mode, an Actor provides a web server which can be used as a website, API, or an MCP server.

Apify vocabulary and the platform model are defined once, in the agent quickstart at https://apify.com/agents.md.

## How to integrate an Actor?

If asked about integration, you help developers integrate Actors into their projects.
You adapt to their stack and deliver integrations that are safe, well-documented, and production-ready.

Do not guess an integration path. Every one of them is in the agent quickstart at https://apify.com/agents.md: the Apify MCP server, Agent Skills with the Apify CLI, the JavaScript and Python clients, the REST API, and the account-free path for an agent with no human to sign in. It also carries the rule on stating cost before the first paid run.

For examples already wired to this Actor's own input schema, see the [API](#api) section below.

Each client library has reference documentation the quickstart does not restate: [JavaScript/TypeScript](https://docs.apify.com/api/client/js/docs.md) (`npm install apify-client`) and [Python](https://docs.apify.com/api/client/python/docs.md) (`pip install apify-client`).

# README

## Web to Markdown Converter for RAG

Convert web pages into clean, token-efficient Markdown optimized for Large Language Models (LLMs) and Retrieval Augmented Generation (RAG) systems with advanced features for production use.

### 🚀 Features

#### Core Extraction

- **Clean Content Extraction**: Uses Mozilla's Readability algorithm to extract only the main article content, removing navigation, footers, ads, and other clutter
- **Markdown Conversion**: Converts HTML to clean, readable Markdown format
- **Smart Link Resolution**: Automatically converts relative URLs to absolute URLs for functional references

#### RAG Optimization

- **Smart Semantic Chunking**: Automatically splits content into embedding-friendly chunks (configurable size, default 1000 chars)
  - Preserves heading hierarchy as context breadcrumbs
  - Configurable overlap between chunks for better continuity
  - Token estimates for each chunk
- **Metadata Extraction**: Captures author, publish date, keywords, description, language, and content type
- **Code Block Detection**: Identifies and extracts code blocks with language detection
  - Separate tracking of fenced code blocks and inline code
  - Line counts and language identification

#### Quality & Deduplication

- **Content Quality Metrics**:
  - Text density, word count, reading time
  - Paragraph and sentence analysis
  - Structure scoring (presence of lists, headings, links)
- **Deduplication Hashing**: SHA256 content hashing and similarity hashing for duplicate detection
- **Production Ready**: Docker-based, robust error handling, deployable to Apify platform

### 📋 Use Cases

- Building RAG knowledge bases from web content with optimal chunk sizes
- Preparing training data for LLMs with quality metrics
- Creating clean documentation from web pages with preserved structure
- Batch processing articles with deduplication
- Web content archival in Markdown format with full metadata

### 🎯 Input

The actor accepts the following input parameters:

#### `start_urls` (required)

Array of URL objects to scrape and convert.

**Example:**

```json
[
  { "url": "https://example.com/article1" },
  { "url": "https://example.com/article2" }
]
```

#### `include_links` (optional, default: `true`)

### 📤 Output

The actor pushes results to the default Apify Dataset. Each scraped URL produces a comprehensive object:

```json
{
  "url": "https://example.com/article",
  "title": "Article Title",
  "markdown_content": "# Article Title\n\nClean markdown content...",
  "chunks": [
    {
      "chunk_id": 1,
      "content": "First semantic chunk...",
      "heading_context": "Introduction > Background",
      "char_count": 850,
      "estimated_tokens": 212
    }
  ],
  "metadata": {
    "url": "https://example.com/article",
    "domain": "example.com",
    "scraped_at": "2025-12-11T01:23:45.123Z",
    "author": "John Doe",
    "publish_date": "2024-01-15",
    "last_modified": "2024-03-20",
    "language": "en",
    "keywords": ["web scraping", "python", "rag"],
    "description": "Article description from meta tags",
    "content_type": "blog"
  },
  "code_blocks": {
    "fenced_blocks": [
      {
        "language": "python",
        "code": "def example():\n    pass",
        "lines": 2
      }
    ],
    "inline_code_count": 15,
    "has_code": true
  },
  "quality_metrics": {
    "text_density": 0.75,
    "paragraph_count": 25,
    "word_count": 1500,
    "sentence_count": 85,
    "avg_sentence_length": 18.5,
    "reading_time_minutes": 8,
    "has_lists": true,
    "has_headings": true,
    "has_links": true,
    "structure_score": 1.0
  },
  "hashes": {
    "content_hash": "a1b2c3d4e5f6...",
    "similarity_hash": "1234567890abcdef"
  },
  "total_chunks": 3,
  "total_chars": 2500,
  "estimated_tokens": 625
}
```

#### Output Fields

**Basic Info:**

- **url**: The original URL that was scraped
- **title**: The extracted page title
- **markdown_content**: The main content converted to Markdown format (with absolute URLs)

**Chunks (for RAG):**

- **chunks**: Array of semantic chunks optimized for embeddings
  - `chunk_id`: Sequential identifier
  - `content`: The chunk text
  - `heading_context`: Breadcrumb trail of headings (e.g., "Chapter 1 > Section 2")
  - `char_count`: Character count
  - `estimated_tokens`: Approximate token count (1 token ≈ 4 chars)

**Metadata:**

- **domain**: Extracted domain name
- **scraped_at**: ISO timestamp of scraping
- **author**: Page author (from meta tags or JSON-LD)
- **publish_date**: Publication date
- **last_modified**: Last modification date
- **language**: Two-letter language code
- **keywords**: Array of keywords from meta tags
- **description**: Page description
- **content_type**: Detected type (blog, documentation, wiki, product, general)

**Code Blocks:**

- **fenced_blocks**: Array of code blocks with language and line count
- **inline_code_count**: Number of inline code snippets
- **has_code**: Boolean indicating presence of code

**Quality Metrics:**

- **text_density**: Ratio of text to HTML (higher = cleaner content)
- **word_count**: Total words
- **reading_time_minutes**: Estimated reading time
- **structure_score**: 0-1 score based on presence of lists, headings, and links

**Deduplication:**

- **content_hash**: SHA256 hash for exact duplicate detection
- **similarity_hash**: Normalized hash for near-duplicate detection

The actor pushes results to the default Apify Dataset. Each scraped URL produces one object:

```json
{
  "url": "https://example.com/article",
  "title": "Article Title",
  "markdown_content": "# Article Title\n\nClean markdown content..."
}
```

#### Output Fields

- **url**: The original URL that was scraped

### 💡 Tips

- **Chunking**: Default chunk size is 1000 characters with 100 character overlap - ideal for most embedding models
- **Rate Limiting**: For large batches, consider adding delays between requests
- **Custom Headers**: The actor includes a User-Agent header to avoid blocks
- **Timeout**: Default timeout is 30 seconds per URL
- **Error Handling**: Failed URLs are logged but don't stop the actor

### 📝 Example Use in RAG Pipeline

````python
## After running the actor, use the output in your RAG system
from apify_client import ApifyClient

client = ApifyClient(token='YOUR_TOKEN')
run = client.actor('YOUR_ACTOR_ID').call(run_input={
    'start_urls': [
        {'url': 'https://docs.example.com/page1'},
        {'url': 'https://docs.example.com/page2'}
    ],
    'include_links': False
})

## Get dataset items with all enhancements
for item in client.dataset(run['defaultDatasetId']).iterate_items():
    # Filter by quality
    if item['quality_metrics']['structure_score'] < 0.5:
        continue  # Skip low-quality content
    
    # Check for duplicates using similarity hash
    if item['hashes']['similarity_hash'] in seen_hashes:
        continue  # Skip duplicate
    seen_hashes.add(item['hashes']['similarity_hash'])
    
    # Process each chunk for embeddings
    for chunk in item['chunks']:
        # Use heading context for better retrieval
        context = f"{item['metadata']['domain']} - {chunk['heading_context']}"
        
        # Embed with your preferred model
        embedding = embed_model.encode(chunk['content'])
        
        # Store in vector database with metadata
        vector_db.upsert(
            id=f"{item['url']}#{chunk['chunk_id']}",
            vector=embedding,
            metadata={
                'url': item['url'],
                'title': item['title'],
                'context': context,
                'content_type': item['metadata']['content_type'],
                'has_code': item['code_blocks']['has_code'],
                'chunk_tokens': chunk['estimated_tokens']
            }
        )
```**Run locally**
```bash
apify run
````

#### Testing Locally

Create a `.actor/INPUT.json` file:

```json
{
  "start_urls": [
    { "url": "https://example.com" }
  ],
  "include_links": true
}
```

Then run:

```bash
python src/main.py
```

### 🚢 Deployment

#### Deploy to Apify

1. **Install Apify CLI**

```bash
npm install -g apify-cli
```

2. **Login to Apify**

```bash
apify login
```

3. **Push to Apify**

```bash
apify push
```

#### Docker Build

To build the Docker image locally:

```bash
docker build -t web-to-markdown .
```

### 📦 Dependencies

- **apify**: Apify SDK for Python
- **requests**: HTTP library for fetching web pages
- **readability-lxml**: Extract main content from HTML
- **markdownify**: Convert HTML to Markdown
- **lxml**: XML/HTML processing library

### 🔧 Configuration

The actor is configured via:

- `.actor/input_schema.json`: Defines the input UI in Apify Console
- `Dockerfile`: Container configuration
- `requirements.txt`: Python dependencies

### 💡 Tips

- **Rate Limiting**: For large batches, consider adding delays between requests
- **Custom Headers**: The actor includes a User-Agent header to avoid blocks
- **Timeout**: Default timeout is 30 seconds per URL
- **Error Handling**: Failed URLs are logged but don't stop the actor

### 📝 Example Use in RAG Pipeline

```python
## After running the actor, use the output in your RAG system
from apify_client import ApifyClient

client = ApifyClient(token='YOUR_TOKEN')
run = client.actor('YOUR_ACTOR_ID').call(run_input={
    'start_urls': [
        {'url': 'https://docs.example.com/page1'},
        {'url': 'https://docs.example.com/page2'}
    ],
    'include_links': False
})

## Get dataset items
for item in client.dataset(run['defaultDatasetId']).iterate_items():
    # Feed markdown_content to your embedding model
    process_for_rag(item['markdown_content'])
```

### 🐛 Troubleshooting

**No output produced:**

- Check that URLs are accessible and return HTML content
- Verify input format matches the schema

**Markdown quality issues:**

- Some pages may have poor HTML structure
- Readability works best with article-style content

**Timeout errors:**

- Increase timeout in `requests.get()` if needed
- Check network connectivity

### 📄 License

This project is open source and available for use in RAG and LLM applications.

### 🤝 Contributing

Contributions are welcome! Please feel free to submit issues or pull requests.

### 📧 Support

For issues or questions, please open a GitHub issue or contact via Apify platform.

# Actor input Schema

## `start_urls` (type: `array`):

List of URLs to scrape and convert to Markdown

## `include_links` (type: `boolean`):

Whether to preserve hyperlinks in the Markdown output. If false, links will be converted to plain text.

## Actor input object example

```json
{
  "start_urls": [
    {
      "url": "https://example.com"
    }
  ],
  "include_links": true
}
```

# API

You can run this Actor programmatically using our API. Below are code examples in JavaScript, Python, and CLI, as well as the OpenAPI specification and MCP server setup.

## JavaScript example

```javascript
import { ApifyClient } from 'apify-client';

// Initialize the ApifyClient with your Apify API token
// Replace the '<YOUR_API_TOKEN>' with your token
const client = new ApifyClient({
    token: '<YOUR_API_TOKEN>',
});

// Prepare Actor input
const input = {
    "start_urls": [
        {
            "url": "https://example.com"
        }
    ]
};

// Run the Actor and wait for it to finish
const run = await client.actor("cooldev/rag-markdown-cleaner").call(input);

// Fetch and print Actor results from the run's dataset (if any)
console.log('Results from dataset');
console.log(`💾 Check your data here: https://console.apify.com/storage/datasets/${run.defaultDatasetId}`);
const { items } = await client.dataset(run.defaultDatasetId).listItems();
items.forEach((item) => {
    console.dir(item);
});

// 📚 Want to learn more 📖? Go to → https://docs.apify.com/api/client/js/docs

```

## Python example

```python
from apify_client import ApifyClient

# Initialize the ApifyClient with your Apify API token
# Replace '<YOUR_API_TOKEN>' with your token.
client = ApifyClient("<YOUR_API_TOKEN>")

# Prepare the Actor input
run_input = { "start_urls": [{ "url": "https://example.com" }] }

# Run the Actor and wait for it to finish
run = client.actor("cooldev/rag-markdown-cleaner").call(run_input=run_input)

# Fetch and print Actor results from the run's dataset (if there are any)
print(f"💾 Check your data here: https://console.apify.com/storage/datasets/{run.default_dataset_id}")
for item in client.dataset(run.default_dataset_id).iterate_items():
    print(item)

# 📚 Want to learn more 📖? Go to → https://docs.apify.com/api/client/python/docs/quick-start

```

## CLI example

```bash
echo '{
  "start_urls": [
    {
      "url": "https://example.com"
    }
  ]
}' |
apify call cooldev/rag-markdown-cleaner --silent --output-dataset

```

## MCP server setup

```json
{
    "mcpServers": {
        "apify": {
            "type": "http",
            "url": "https://mcp.apify.com/?tools=fetch-actor-details,cooldev/rag-markdown-cleaner"
        }
    }
}
```

The hosted server signs you in with OAuth on first connect, so no API token belongs in this config. Clients without OAuth support can send an `Authorization: Bearer <APIFY_API_TOKEN>` header instead, using a token from API & Integrations in Apify Console (https://console.apify.com/settings/integrations).

## OpenAPI specification

Download the OpenAPI definition: https://api.apify.com/v2/actors/TqxgZh8FIJdikFhGP/builds/aDofwbpjFLL6qujF9/openapi.json