ContentForge

AI-powered SEO content generator with parallel processing. Generate blog articles, comparison tables, and optimized content at scale.

---

Features · Installation · Configuration · Usage · Supported Models · Cost Estimation · License

---

Features

• Parallel Generation - All content sections generated simultaneously using asyncio
• SEO Optimization - Keyword targeting, density control, E-E-A-T compliance
• Comparison Tables - Dynamic tables with customizable columns and star ratings
• Dual Output - Generate both HTML and Markdown formats
• YAML Configuration - Simple template-based setup, no coding required
• Preview Mode - See structure and cost estimates before generation
• Multi-language - Generate content in any language
• Retry Mechanism - Automatic retry with exponential backoff

Installation

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

# Install dependencies
pip install -r requirements.txt

# Copy environment template and add your API key
cp .env.example .env

Configuration

Edit config.yaml to customize your content:

# Page title
title: "Best Web Hosting Providers 2025"

# SEO settings
seo:
  primary_keyword: "web hosting"
  secondary_keywords:
    - "VPS hosting"
    - "shared hosting"
  keyword_density: 2.0
  tone: "informative"

# Content sections
sections:
  - heading: "What is Shared Hosting?"
    words: 120
  - heading: "VPS vs Dedicated Server"
    words: 120

# Comparison table
table:
  enabled: true
  rows: 5
  columns:
    - name: "item"
      header: "Provider"
    - name: "rating"
      header: "Rating"
      type: "stars"

# AI model
model: "openai/gpt-4o-mini"

# Output format: html or md
output: "md"

Usage

# Generate content
python main.py

# Preview mode (no API calls)
python main.py -p

# Use custom config file
python main.py -c custom_config.yaml

# Create example config
python main.py --init

Environment Variables

Create a .env file with your OpenRouter API key:

OPENROUTER_API_KEY=your_api_key_here

Get your API key from OpenRouter.

Project Structure

contentforge/
├── main.py              # Entry point
├── config.yaml          # Configuration template
├── requirements.txt     # Dependencies
├── api/
│   └── client.py        # OpenRouter API client
├── core/
│   ├── config.py        # Config loader
│   ├── constants.py     # Defaults and prompts
│   ├── exceptions.py    # Custom exceptions
│   └── models.py        # Data models
├── generators/
│   ├── html.py          # HTML output builder
│   ├── markdown.py      # Markdown output builder
│   └── prompts.py       # AI prompt templates
├── utils/
│   └── text.py          # Text utilities
└── content/             # Generated output

Supported Models

Model	Speed	Cost/Section	Best For
openai/gpt-4o-mini	~45 tps	~$0.00012	Bulk content, drafts
openai/gpt-4o	~30 tps	~$0.003	High-quality articles
openai/gpt-oss-20b:free	~40 tps	Free	Long content, testing

Using Free Models

To use free models (e.g., gpt-oss-20b:free), you need to enable data training in your OpenRouter account:

1. Go to OpenRouter Privacy Settings
2. Enable "Enable free endpoints that may train on inputs"

Note: Free models may use your prompts and completions for training purposes.

Cost Estimation

Based on real usage data with GPT-4o-mini:

Metric	Value
Avg input tokens	~340
Avg output tokens	~120
Cost per API call	~$0.00012

Calculate Your Cost

A typical article has:

• 1 introduction
• 1 comparison table (optional)
• 3-5 content sections
• 1 conclusion

Total: 5-8 API calls per article

Articles	API Calls	Estimated Cost
1	6	$0.0007
10	60	$0.007
100	600	$0.07
1,000	6,000	$0.72
10,000	60,000	$7.20

Use python main.py -p (preview mode) to see exact API call count and cost estimate before generation.

License

MIT License - see LICENSE for details.