Words belong on the beat.
Forced alignment works from supplied lyrics to estimate word timing. The toolkit also has a transcription fallback for material without prepared lyrics.
CASE STUDY / VIDEO TOOLKITFROM AUDIO TO THE TIMELINE
Video Toolkit
A set of local tools for the repetitive side of music videos: align lyrics, prepare subtitles, place markers and get the project ready for DaVinci Resolve.

INSIDE THE TOOLBOX
The actual Toolbox in Video Toolkit: prepare the song, work on the video, then check it before upload.
Explore the full interface ↗Focused views of the app’s existing Toolbox interface, with original screenshots available below each view. No media files are loaded. The interface is read-only; tool details can be opened, and no processing jobs run from this portfolio.
THE STARTING POINT
Making a music video meant doing the same preparation again: build the folders, line up the words, mark the beats and make sure the media would behave in the editor.
Video Toolkit gathers that preparation into a local workflow. Python handles the media jobs, Lua and the Resolve API connect to the edit, and ffmpeg does the format work.
HOW IT FITS TOGETHER
Prepare the files and timing before making the creative choices in the edit.
Organize the audio and project folders before the edit.
LOCAL FILESAlign supplied lyrics to vocals and create subtitles and beat markers.
ALIGNMENT · MARKERSBring the prepared material into the timeline and check the export.
RESOLVE · FFMPEGINSIDE THE BUILD
Forced alignment works from supplied lyrics to estimate word timing. The toolkit also has a transcription fallback for material without prepared lyrics.
Project folders, subtitle files, markers and checklists give the next editing session a consistent starting point.
Vertical clip preparation, caption retiming, still extraction and codec conversion cover other recurring tasks in the same workflow.
Automate the repeat.
Keep the decisions.
WHAT I TOOK FROM IT
The interesting work is in the edit. This toolkit exists because I kept noticing preparation I could describe clearly enough to automate.
NEXT PROJECT / ATLAS FETCH