From wiring nodes by hand to creative programming with @ mentions

Here is how reusing a reference image used to feel. You have a character portrait sitting in one corner of your canvas and the shot you want to generate way over in another. To connect them, you find the portrait, drag a reference edge out of its image, pan across the canvas, and drop it onto the node you are working on. Want that same portrait in five different shots? You wire it five times. Want a whole mood board as reference? You drag an edge for every single image.. and then do it all again every time the board grows.

It works, but it is fiddly and spatial. You end up spending more time navigating the canvas and untangling connections than actually creating. The board turns into a web of wires, and the idea in your head gets lost in the plumbing.

The @ mention replaces all of that. Instead of dragging edges between boxes, you just name what you want, right inside your prompt.

It is closer to creative programming

Think of a mention as a variable. In code, you do not paste the same value everywhere.. you name it once and refer to it by that name wherever you need it. @ mentions work the same way for creative work. You point at something by name, and Vilva fills in the actual image, text, or audio behind it.

That changes how it feels to build. You are no longer wiring a diagram. You are composing with named pieces.. "use this character here, that mood board there, this voice for the narration".. and every reference turns into the real thing when you generate. Change the source and every mention of it updates. Reuse a reference in ten prompts by simply mentioning it ten times. No edges, no hunting, no copies to keep in sync.

The rest of this post is the quick tour of how it works.

Just type @

Type @ anywhere in a prompt and a small menu pops up with everything on your canvas. Pick the one you want and Vilva drops it in as a neat little tag. You can hover that tag to see what is inside, and it never turns into a messy link. You never drag an edge, copy anything, or type anything technical.. Vilva quietly remembers exactly which item you meant, even when you have a dozen things with similar names.

Pick exactly the part you want

Once you have pointed at something, you can bring along the whole thing, or just one part of it. Each part has a short name you add after a dot:

.title .. just its title
.summary .. the short description
.prompt .. its main text
.image1, .image2, .image3 .. a specific image on it (the first, the second, the third..)
.audio1, .audio2 .. a specific audio clip
.video1 .. a specific video

So @Hero.image1 means "the first image on my Hero card," and @Narration.audio1 means "the first audio clip on Narration." You do not have to memorize any of this. The menu lists the choices for you and fills in the rest. The short names are just handy to recognize when you see them in a tag. This is the "variable" idea in practice.. a name that stands in for the real content.

Grab a whole group at once: `.images`

When you want every image on something, or (more usefully) every image inside a group like a mood board, picking them one at a time (.image1, .image2, .image3..) gets tedious. And, just like the old edge-dragging, you would have to remember to come back and add the new ones every time the board grows.

The plural shortcut does it for you:

@Mood Board.images

That pulls in every image in the group at once, in order. Add a new image to the board later and it is automatically included the next time you generate. The same idea works for the rest: .audios for all the audio, .videos for all the video, and .files for all the files.

If a group is still empty, the .images option still shows up (just greyed out) so you always know where to find it once you start adding things.

Reference the same image twice? No problem

Sometimes you will point at the same image in two ways without realizing it. Once on its own with .image1, and once because it also lives inside a mood board you tagged with .images. You might expect that to send the picture twice and muddle the result.

Vilva handles it for you. Within a single prompt, the same image only counts once, no matter how many ways you have referenced it. Your results stay clean and you never have to think about it.

A quick example

Imagine you have:

a Hero card with one image, a portrait of Harry,
and a Mood Board group with three images: that same portrait, a neon skyline, and a wet street.

The old way, you would drag four reference edges across the canvas, one per image, and rewire them whenever anything changed. Now you just write:

Wide shot of @Hero.image1 against @Mood Board.images, moody noir tone.

Vilva pulls in Harry's portrait and all three mood board images, notices the portrait shows up in both places, and only uses it once. You get the shot you pictured, without the clutter, and without ever leaving the prompt.

The short version

The old way: drag a reference edge from each image, across the canvas, for every node.. and redo it whenever something changes.
The new way: type @, name what you want, and Vilva fills in the real image. References behave like variables you can reuse anywhere.
Add a dot to grab one part (.image1, .audio1, .title..), or a plural shortcut (.images, .audios, .videos, .files) to pull in a whole mood board at once.
Reference the same image more than once and Vilva keeps it clean automatically.

It all works the same whether you are writing the prompt yourself or asking Vilva's AI to do it for you. The point is simple: stop wiring boxes together, and start doing creative programming.