Make a video, start to finish
A worked example. Take a 30-second script and finish it in the editor, with the options explained at each step.
On this page
Start with the script#
This walkthrough follows one short promo from a written script to an exported file. Follow along with your own words and footage. Every step names your options, so you can choose differently where your edit needs it.
PROMO / 30 SECONDS
VO Every team says they move fast.
Then the handoffs start.
A brief here. Feedback there.
Three tools before anyone sees a cut.
Northwind puts it in one place.
Write, review, and ship from the same board.
Northwind. Ship the work, not the process.
SUPER at 0:26 northwind.comThe voiceover sets the timing for everything else
When narration leads the video, cut the voiceover first. Get its timing right before you place a single shot, and the picture follows easily. This walkthrough is ordered that way on purpose.
Create the project#
This happens in VideoDraft Editor, not the ADE window. Open it from Command Center, or from the Editor tab in the right panel of your task. Then choose File > New Project.
New Project asks for a name, a destination, an aspect ratio, a resolution, and a frame rate. These set the canvas every clip is fitted to, so take a moment now.
| Option | How to choose |
|---|---|
| Aspect ratio | 16:9 for web and YouTube. 9:16 for Reels, TikTok, and Shorts. 1:1 for feed posts. 2.4:1 for a widescreen look. |
| Resolution | 1080p covers most delivery. Pick 4K when you plan to punch in or reframe, since cropping a 4K source into a 1080p canvas keeps full detail. |
| Frame rate | Match your footage. 24 fps reads filmic, 30 fps is the safe web default, and 60 fps holds up for fast motion. |
For this promo: 16:9, 1080p, 30 fps. The project saves as a .vdproject package: one folder holding the timeline, the media list, and generated media. The editor writes each edit to disk within milliseconds, so a crash costs you the current move and nothing more.
Bring in your media#
Press Cmd + I, or drag files onto the Production Drawer, the panel on the left. Its menu holds Media, Colors, Text, Effects, and Transitions, and audio sits under Media with video and stills. You never import captions, because the editor builds them from the speech already on the sequence.
- The voiceover recording, as a single audio file.
- Four or five B-roll shots, the cutaway footage that covers the voiceover. Bring in more than you need and trim them down later.
- A logo or end card still, if you have one.
- A music bed, at least as long as the finished piece.
A generated voiceover works for timing the edit
You do not need the final recording yet. The bundled videodraft CLI makes voice, music, and sound effects from a prompt, and the result imports like any other file. Use it to time the edit before the real read exists.
Lay down the voiceover#
Drag the voiceover onto an audio track at frame 0. That single clip is now the length of your video, and every later decision is measured against it.
Tracks are layers. Index 0 renders on top, video and images share video tracks, and audio needs an audio track of its own. A video clip that carries sound arrives with a linked audio clip that stays in sync through trims and speed changes.
Tighten the voiceover#
This step saves the most time. The editor transcribes the timeline on your Mac, so you cut by the words rather than by the shape of the audio.
- 1Remove dead air first, meaning the quiet stretches with no speech. Silence trimming finds them, deletes them, and slides everything after each cut earlier to close the gap.
- 2Then cut by word. Remove filler words everywhere they occur, or select a fumbled phrase in the transcript and delete it.
- 3Check each cut. Tight pulls the neighbors close and can sound clipped, balanced keeps a natural beat, and loose leaves breathing room.
Dynamic music can read as dead air
Detection measures each stretch against how loud the speech in that recording is. Steady music or room tone is left alone, but a quiet intro or a fade can read as dead air. Transcribe musical sources first, or check the result before keeping it.
Cover the voiceover with shots#
With the voiceover timed, place B-roll on a video track above it. Work down the script line by line. Each sentence gets a shot.
| Key | What it does |
|---|---|
| Cmd + B | Split a clip at the playhead |
| Q or W | Trim the start or end to the playhead |
| Backspace | Delete and close the gap |
| Opt + Backspace | Delete but hold the slot open |
| Cmd + T | Add a cross dissolve at a cut |
| + / - | Zoom the timeline |
Backspace closes the gap, so every clip after it slides earlier. Use Opt + Backspace to hold the slot open for a shot that has not arrived yet. Trim edges drag outward as well as inward, so a clip can always go back to its full source length.
- A shot runs long. Set a constant speed from 0.01x to 1000x, or build a ramp, meaning a speed that changes across the clip.
- Two shots need to blend. A transition attaches to the incoming clip, the second of the pair. Cross dissolve is the default, and dip to black, blur dissolve, whip pan, and cross zoom are there too.
- Two shots need to share the frame. Use a split screen, grid, or picture-in-picture layout instead of positioning them by hand.
Captions and the end card#
Captions come from the speech already on the timeline. You do not select a clip first. The editor finds the spoken content, transcribes it, and builds styled caption clips on a track of their own.
| Option | What it changes |
|---|---|
| Words per caption | 3 to 5 words reads well on a phone. Higher counts suit wider screens. |
| Highlight color | Tints the word currently being spoken. |
| Animation | Word-level presets such as word reveal, word pop, and highlight block, plus simpler fades and slides. |
| Position | A vertical center around 0.8 sits captions low without clipping the frame edge. |
| Casing and profanity | Force upper or lower case, and mask profanity if the cut needs it. |
The script calls for on-screen text at 0:26, which a script marks as a SUPER. Add it as a title over the closing shot, across exactly the frames you want it visible. Titles carry font, size, color, outline, background fill, and the same animation presets.
Grade, mix, and export#
- 1Grade the picture, meaning match the color of your shots to each other. Measure with scopes, compare against a reference shot, adjust, and repeat until they match.
- 2Put the music on a track under the voiceover, then duck it, meaning lower the music while the voice speaks. Cmd + click the music clip to add a volume keyframe. Pull the level down where the voice comes in, and back up in the gaps.
- 3Watch it once, end to end, without touching anything.
- 4Export.
Or hand the whole thing to an agent#
Every step above is available to an agent through the same project, undo, and export paths. You can delegate the whole job, or take over at any point, because you are both editing one project rather than trading files.
Build a 16:9 1080p 30 fps promo from promo.txt and the media in this folder.
Put the voiceover on an audio track, trim dead air, and remove filler words.
Cover each line of the script with a B-roll shot, cutting on the voiceover.
Generate captions, 4 words each, low in frame, with the active word highlighted.
Add "northwind.com" as on-screen text at 0:26 over the closing shot.
Grade the B-roll to match, duck the music under the voice, and export H.264 1080p.The editor window updates while the agent works
Changes appear live, so there is nothing to refresh. Open the editor whenever you want to see the current state or take the edit back.