Short-form video is the highest-ROI content format right now, but making it is still surprisingly manual. Even with AI tools, you're still stitching things together yourself. I wanted to see if the entire pipeline — script, voice, visuals, captions — could be fully automated end-to-end. Turns out it can. Clippie uses Claude for script generation, ElevenLabs for voice narration, and Remotion for programmatic video rendering with animated captions and scene transitions. The whole thing runs serverless — no FFmpeg hacks, no manual steps. The process evolved from "can this even work?" to obsessing over output quality. Early versions were technically functional but felt like AI slop. The unlock was treating it like a real production pipeline: proper voice pacing, music ducking, smooth transitions, and word-level caption timing. That's what makes people actually want to post the output.

Clippie.online

Turn any idea into a viral video.

Turn any idea into a viral video.