LLM Content Proxy

A Python package for extracting the main content from web pages. Can be used as a standalone server or deployed as a cloud function (GCP, AWS or Azure).

Primary purpose was to proxify blog articles content to LLM for summarizing or analysis.

Features

• Extract the main textual content from web pages
• Remove unwanted elements like ads, navigation, headers, etc.
• Simple REST API to access this functionality
• Multiple deployment options:
  • Standalone FastAPI server
  • Google Cloud Function
  • AWS Lambda
  • Azure Functions

Installation

# Clone the repository
git clone https://github.com/yourusername/llm_content_proxy.git
cd llm_content_proxy

# Install the base package
pip install -e .

# For GCP development
pip install -e ".[gcp]"

# For AWS development
pip install -e ".[aws]"

# For AWS development
pip install -e ".[azure]"

# For development tools (testing, linting)
pip install -e ".[dev]"

Usage

As a Python Package

from llm_content_proxy import ContentExtractor

# Create an extractor instance
extractor = ContentExtractor()

# Extract content from a URL
result = extractor.extract_from_url("https://example.com/article")

# Access the extracted content
print(f"Title: {result['title']}")
print(f"Content: {result['content']}")
print(f"Word count: {result['word_count']}")

Running the Standalone Server

# Method 1: Using the command-line entry point
llm-content-proxy

# Method 2: Using Python directly
python -m server.app

Then access the API at: http://localhost:8000/?link=https://example.com

Deployment Options

This package supports multiple deployment options:

1. Standalone Server: Run as a FastAPI application on your own server
2. Google Cloud Functions: Deploy as a serverless function on GCP
3. AWS Lambda: Deploy as a serverless function on AWS
4. Azure Functions: Deploy as a serverless function on Azure

See the setup documentation for each option:

• Standalone Server Setup
• Google Cloud Functions Setup
• AWS Lambda Setup
• Azure Functions Setup

API Usage

Once deployed, you can access the API at:

GET /?link=https://example.com/article

The response will be a JSON object with the following structure:

{
  "title": "Article Title",
  "content": "The main content of the article...",
  "url": "https://example.com/article",
  "word_count": 1234
}

Example LLM Integration

To use this with an LLM, you can format your prompts like:

Summarize the following article from {url}:

{content}

Configuration

The server listens on all interfaces (0.0.0.0) on port 8000 by default. You can modify this when running the server:

from llm_content_proxy import run_server

# Run on a different host and port
run_server(host="127.0.0.1", port=5000)

Requirements

• Python 3.7+
• Common requirements:
  • requests
  • beautifulsoup4
• Standalone server:
  • fastapi
  • uvicorn
  • pydantic
• GCP:
  • functions-framework
• AWS:
  • boto3 (for deployment)

License

WTF