| Hello HackerNews, I built a self-hosted boilerplate for blog/content creation because I think current AI writers have a big red flag: their limitations. My ideal AI writer should work for multiple projects and generate as many articles as needed. How it works:
First, you create a project and add topics. Each project represents a complete blog page, you can customize metadata, header, footer, CTA, and social media links. I spend some time on how to generate content at best and I am happy with final result, here how it creates the content: - Find relevant articles about your topic using advanced search algorithms - Analyze and score these articles for quality (humanization, technical accuracy, and readability) - Choose which articles to use right from the UI with an intuitive selection interface - Pick from 6 different AI models (Claude, Gemini, OpenAI, XAI, Cohere, and Mistral) to create an article with references to those relevant articles—this way, we eliminate hallucination risk - OpenAI scoring: After article creation, OpenAI scores the generated article based on quality metrics - Iterate and publish: Create infinite versions of an article, pick the best one, and publish - You can edit everything about your generated posts including content, meta tags, and URL slugs to perfect your SEO and content strategy Looking forward to hear your feedback and opinions, cheers! |