One toolkit to automate an entire motion design pipeline.

Loading video
MDKit v6 release video

Context

MDKit is the internal production system used every day by Vidsy's team of more than twenty motion designers. It connects live briefs, shared project files in Dropbox and After Effects, automating the workflow from setup and ingest through rendering and delivery.

Measured from its own usage analytics, it is on track to save 8,000 to 10,000 production hours a year.

At that scale, small tasks quickly multiply across creators, formats and languages. I built the first version in 2022 to remove that repetition and give the whole team a faster, more consistent way to work.

The native macOS app's Studio tab: audio, visuals, text and utility tools for any phase of production.
The native macOS app's Studio tab: audio, visuals, text and utility tools for any phase of production.

Role

  • Workflow and product design
  • Native macOS and full-stack development
  • After Effects automation, platform and AI integration
  • Rollout, documentation and ongoing iteration
The After Effects extension's CGC Auto tab: drop in footage, then safe zones, audio and scene detection run on build.
The After Effects extension's CGC Auto tab: drop in footage, then safe zones, audio and scene detection run on build.

Solution

  • Starts from a live brief in Vidsy's production platform, pulling in creators, deliverables, languages and formats.
  • Finds and organises footage, audio, scripts and brand assets in Dropbox.
  • Turns creator footage into structured After Effects projects, using fast native cut detection to prepare sequences and audio.
  • Automates composition setup, naming, safe zones, format adaptation, localisation and high-volume iteration.
  • Runs on-device Apple Intelligence for transcription, captions, OCR, and subject and face detection, with small foundation models sorting and sense-checking text and spotting patterns in client feedback.
  • Calls external models through their APIs, such as ElevenLabs for voice, then carries completed work through rendering, upload and archiving.
Render and Upload in the native app: format, pre-upload glitch check, platform upload and Slack notification.
Render and Upload in the native app: format, pre-upload glitch check, platform upload and Slack notification.

Impact

MDKit has made production faster and more consistent, removing error-prone manual work while leaving creative decisions with the designer.

It has grown with the team since 2022: from a ScriptUI panel, to a JavaScript CEP extension, to a React and TypeScript extension with connected services and AI-assisted tools, and now to a native SwiftUI macOS app.

Each stage expanded the same product into another part of the workflow. Today, MDKit connects Vidsy's platform for briefs and delivery, Dropbox, After Effects, and on-device and cloud AI models as one system.

Related Work