OPEN-SOURCE AI PIPELINE FOR AUTOMATED SHORT-FORM VIDEO CREATION: A PROOF-OF-CONCEPT STUDY
OPEN-SOURCE AI PIPELINE FOR AUTOMATED SHORT-FORM VIDEO CREATION: A PROOF-OF-CONCEPT STUDY

Summary/Abstract: Abstract: The proliferation of short-form video platforms creates unprecedented demand for content automation. Manual production requires 2-4 hours per video, while commercial platforms impose subscription costs of €20-47 monthly. This study presents ClipClap Factory, an open-source automation pipeline employing n8n workflow orchestration and local LLM inference (Google Gemma 3 12b). Proof-of-concept deployment on Instagram validated the system through 112 workflow executions during summer 2025. Key findings demonstrate: (1) 67% success rate with 75 published videos; (2) €0.35 per video production cost representing 99% cost reduction versus manual production; (3) 4-7 minute production velocity; (4) 5-second square-format videos (1024×1024) optimized for Instagram. Audio truncation occurred in 15% of videos when narration exceeded the 5-second Freepik API constraint. Visual-semantic misalignment appeared in 20% due to unpredictable AI generation behavior. The deployment prioritized technical validation over audience optimization; engagement metrics (peak: 383 views, 21 followers) represent incidental outcomes. The hybrid architecture (local LLM with selective commercial API integration) establishes a cost-optimized pattern for proof-of-concept workflow validation. Future work includes script length optimization and multiplatform expansion.

• Details
• Contents