Video editor nodes

1. + on the bottom toolbar → Smart edit (it carries a Beta tag).
2. Drag the handle of a video node onto it. Connect as many as you like — the order you connect them is the order on the timeline.
3. The card goes from Nothing here yet, connect a video node first to Open smart edit. Click that to enter the editing bay.

Wiring isn't a one-shot thing: connect another one later and it's appended to the end of the timeline; disconnect one and that stretch drops off the timeline, with the rest closing up. Any order you set by hand on the timeline survives that sync.

Selecting the card also pops up an input bar along the bottom: describe the cut you want, pick a mode (Default / Explainer / Ad batch / Talking head / Footage mashup) at the bottom left, and the button next to it holds aspect · length · resolution. Sending does not render anything — it hands the footage plus your line to the agent to cut for you, and the finished film lands as a new node; the button on the card remains the do-it-yourself path.

Smart Edit accepts more than video. Select the compose node, wire or drag in what you want to use, then add one line of direction:

• Video determines the timeline material: which clips are available, their order, and the picture and sound each one brings.
• Images can reference a title card, color palette, packaging style or the composition of a particular shot; they do not need to move.
• Text or a script can carry narration, shot order, the lines to keep, subtitle conventions and rhythm.
• Audio lets the finished cut follow an existing voice-over or music track instead of creating a mismatched new one.

The input bar shows footage cards and text cards. Click a card to find where it came from and confirm the Agent is seeing exactly what you intended. You can also choose only a mode and output specification and let the Agent cut from the material; the result lands as a new node, while the original videos, images and text remain untouched.

The editor node generates nothing on its own. What it stores is an edit document: which stretch of which source clip, how fast it plays, how loud it is, what transition sits at the joins, what the subtitles say.

Two things follow from that:

• Not one byte of the source is touched. Shortened, slowed, muted — all of that is only "how it's used this time". You can pull it back out whenever, and it doesn't affect how that video node is used anywhere else.
• Close it and reopen it and it's still there. The edit document lives on the node and is saved with the canvas. Open the same project on another machine and a half-finished cut is waiting for you.

Export, top right in the editing bay, has two destinations:

• Export to file — downloads an mp4.
• Export to canvas — the finished film lands back on the canvas as a new video node. It's no different from any other video node: you can upscale it, remove subtitles, grab frames, or wire it into another editor node.

Details in 7.5.