All Projects Automation · Content Pipeline

RAGE

An automated post generator for content creation platforms. RAGE discovers topics, researches them, generates captions and imagery, and publishes finished posts ready for distribution through the web UI or API.

RAGE — automated content generation pipeline interface
Role Full-Stack Developer
Year 2025
Type Automation Tool
Source GitHub

RAGE is the engine behind automated content creation. It operates as a seven-stage pipeline that takes a single trigger — a POST request to the server — and produces a fully finished social media post without human intervention. Each stage is modular, testable, and handles one specific responsibility.

Stage 1 discovers candidate topics by scraping daily facts from The Fact Site, parsing HTML paragraphs and filtering out boilerplate content. Stage 2 removes previously used topics and uses an LLM to rank the remaining candidates for engagement potential, selecting the most compelling fact. Stage 3 deep-researches the chosen topic via DuckDuckGo, scraping multiple sources and producing an AI-generated summary.

Stage 4 writes a character-driven meme caption using a configured LLM, blending dark humor with the research context. Stage 5 generates an AI image prompt and sends it to the FLUX image generation API to create a custom background. Stage 6 uses Pillow to composite the caption text over the generated image with professional typography and fade effects. Finally, Stage 7 saves the completed post as a PNG and records all metadata.

The pipeline is designed for reliability — each stage validates its inputs and can fail gracefully without corrupting downstream stages. The used-topics registry ensures no fact is ever published twice. RAGE served as the foundation for CURIO's content engine, proving the pipeline architecture before it was adapted for multi-user SaaS deployment.

← Previous GeepSeek Next → CURIO