Case Studies

Problem

Long-form source material (talking-head, tutorial, interview footage) needs to become publishable vertical short-form video, and manual editing takes hours per clip.

Approach

Local Whisper transcription → AI selects the strongest segments using proven hook formulas (contrarian takes, curiosity gaps, emotional beats, numbered lists...) → ffmpeg auto-crops to 9:16, burns word-by-word animated captions with keyword highlighting, adds an intro sound cue, and normalizes loudness to -14 LUFS.

Problem

Long-form videos also need basic post-production (removing filler words, dead air, noise) before they're publish-ready.

Approach

The same automation pipeline in long-form mode, batch-processing filler-word and silence removal.

Problem

Some clients want an upgraded version with transition effects, not the base short-form output.

Approach

An effects/transitions layer added on top of the base pipeline, producing a premium-tier output.

Problem

Finished content still needs to be manually uploaded to a website — another repetitive step.

Approach

Automated publishing of finished content directly to WordPress, chained onto the content pipeline.

Problem

Manually searching Upwork, Reddit, and RemoteOK for suitable freelance opportunities every day is time-consuming and leads get missed.

Approach

A scheduled scan across multiple job sources → AI (Claude) scores each posting against skills/preferences → qualifying leads get written into a Notion tracking database.