Back to selected work
Looking for a similar build? Explore my web, app, and software development services.
AI Automation/2024
Local AI Content Automation Pipeline
End-to-end AI content generation system
A complete local AI pipeline for automated content generation using n8n, Ollama, ComfyUI, and FLUX.
Problem
Creating an automated content pipeline that runs entirely locally with AI-generated visuals and voiceovers.
Solution
Built an end-to-end pipeline using n8n for orchestration, Ollama for text generation, ComfyUI/FLUX for visuals, and Kokoro TTS for voice.
Architecture
n8n orchestration → Ollama text generation → Python scripts → ComfyUI/FLUX visuals → Kokoro TTS voice → Video publishing
Challenges
- — Temporal inconsistencies in video generation
- — Flickering in AI-generated visuals
- — Prompt drift across pipeline stages
- — Coordinating multiple AI systems
Results
- — Fully local content generation pipeline
- — High-quality AI visuals and voice
- — Automated publishing workflow
- — Debugged complex AI pipeline issues
Lessons learned
- — AI pipelines require extensive debugging
- — Temporal consistency is challenging in AI video
- — Prompt engineering affects all stages
- — Local AI enables complete control