Auto Blogs - AI-Powered Content Generation Toolkit

AutoBlogs is an AI-assisted content generation tool that can leverage both open-source and/or proprietary Large Language Model (LLM) agents to create high-quality, SEO-optimized content for various needs. The system integrates a human-in-the-loop workflow, ensuring that AI-generated drafts can be reviewed, edited, and refined before publishing.

Project Overview

This project aims to simplify and accelerate the process of writing contents (blog, articles, book, research papers, etc.) by combining AI automation with human editorial control. It provides a flexible framework where multiple LLM providers can be integrated to create and customize contents using a streamlit dashboard.

The contents can then be published online, like in GitHub Pages, substack, etc. and thus starting a technical writing business or freelance job becomes very easy.

Getting Started

The repository uses third-party Python SDKs that provide an API interface to interact with open-source or any proprietary LLM, like OpenAI or Anthropic Claude to generate content.

The package is available on PyPI and can be installed as follows:

pip install autoblogs

The package does not have a hard dependency on a third-party LLM SDK, but requires one based on the type of model you want to use. For example, it either requires the Anthropic SDK for the CLAUDE provider or the OpenAI SDK for all other types of providers.

Environment Variables

Copy .env.example to .env and fill in your values, or supply them as system environment variables. The available variables are described below.

# LLM provider name - e.g. "OPENAI", "ANTHROPIC", "LOCAL", "NVIDIA-NIM"
LLM_PROVIDER = "LOCAL"

# Model identifier and API key for the selected provider
LLM_MODEL_NAME = "awesome-model"
LLM_MODEL_APIKEY = "your-api-key-here"

# Optional: override the default API base URL (useful for self-hosted or custom endpoints)
LLM_API_BASE_URL = "https://example.com/api/v1"

# Model generation parameters
MAX_TOKENS = 4096
TEMPERATURE = 0.7

# I/O configuration
CONTENT_OUTPUT_DIR = "output"
AGENT_PROMPT_CONTEXT = "base.txt.jinja"

Command Line Tools

The package provides two distinct tools for different workflows. autoblogs-cli is a lightweight interactive terminal tool for quick content generation, while autoblogs-ui launches a full browser-based dashboard for an end-to-end editorial workflow — creation, editing, preview, and file management.

AutoBlogs-CLI

autoblogs-cli runs entirely in the terminal. It reads your configuration from the .env file (or system environment), then walks you through a short interactive session to collect the content requirements before calling the LLM.

autoblogs-cli

Once launched, it prompts for the following inputs in order:

A sample session looks like:

Set the Content Topic (one-line): Linear Regression for Beginners
Explain the Prompt to Generate Content: Explain with real-world examples, keep it beginner-friendly
Set SEO Keywords (comma separated): statistics, regression, machine learning
Input Output Filename: linear-regression.md

The generated post is written to <CONTENT_OUTPUT_DIR>/<Output Filename> (default: output/linear-regression.md).

AutoBlogs-UI

autoblogs-ui launches a multi-page Streamlit dashboard in your browser. It provides the same content generation capabilities as the CLI, plus an inline draft editor, live markdown preview, and file download.

# Launch on the default port (http://localhost:8501)
autoblogs-ui

# Run on a custom port
autoblogs-ui --server.port 8080

# Run in headless / server mode (no browser auto-open)
autoblogs-ui --server.headless true

Any native Streamlit CLI flag is accepted and forwarded directly. Once running, open http://localhost:8501 (or the configured port) in your browser.

The dashboard is organised into five pages:

Contribution Guidelines

All contributions, bug reports, bug fixes, documentation improvements, enhancements, and ideas are welcome. A detailed overview of how to contribute can be found in the contributing guidelines. If you run into an issue, please file a new issue for discussion. Create a pull request for a new feature or a fix to an existing issue here.

As contributors and maintainers to this project, you are expected to abide by PyUtility's code of conduct. More information can be found at: Contributor Code of Conduct.