What is Location Scout?

A MITO feature that takes one image of a location and generates four more views of the same space. You keep all five on the location Element and use them as references, so the room stays the same room across every shot you generate.

Why do my AI backgrounds change between shots?

Because one reference image only shows the model one side of the space. When you ask for a reverse shot or a wide, everything outside that original frame has to be invented, and it gets invented differently each time. That is why shelves move, props vanish and the light shifts between cuts.

How does Location Scout keep the views consistent?

It does not generate four separate images. It runs a five-second camera move around your starting image and takes one frame at each stop, so every view comes out of a single continuous shot of one space. The camera changes angle but holds its height and distance, which is what keeps the room feeling like the same room.

How do I use Location Scout?

Two ways. Ask Director to add more reference points to a location you have already made, or open it yourself from the plus icon in the toolbar, describe the location, attach your existing reference from the canvas, toggle on the four extra views and hit Generate.

Do I need an existing image first?

You need a starting image, but you do not need to have made it elsewhere. Ask Director for the location and it builds one, then run Location Scout on it in the same conversation.

What is the difference between a location Element and Location Scout?

The Element is the thing that holds the location and gets reused across your project. Location Scout is what fills it out, taking it from one angle to four so there is more for the model to work from.

Does this work for outdoor locations?

Yes. The camera move works the same way on a terrace, a hillside or a forest clearing as it does on a bar interior. Open spaces benefit for the same reason interiors do, because a single frame of a landscape still leaves most of it undescribed.

Where can I find the tutorial inside MITO?

Under Tutorials in the sidebar, alongside the rest of them, so you can watch it without leaving the project you are working in.

Introducing Location Scout: Keep Every Shot in the Same Place

Location Scout takes one image of a location and gives you four more views of the same place.

That is the whole feature, and it fixes a problem that shows up the moment a project needs more than one shot in the same room.

Why one image of a location is not enough

A reference image shows the model one side of a space. One wall, one angle, one slice of the room.

Then you cut. A reverse shot, a wide, a close-up from across the room. Everything that was behind the camera in your reference now has to be in frame, and none of it exists yet, so the model invents it. Ask again and it invents something else.

This is where continuity quietly falls apart. The shelves move. A doorway turns up on a different wall. The evening light you set in shot one is daylight by shot four. Each image is fine on its own, and together they are clearly not the same place.

Giving the model several views of the location instead of one pins down the layout, the lighting and the details that a single frame could never describe.

Two ways to use it

Both routes end to end: Director building the location and scouting it, then the same job done by hand.

This walkthrough and the rest of our tutorials are inside MITO too, under Tutorials in the sidebar.

Through Director. Ask for the location and it builds one. A warm, cosy bar with shelves of liquor is enough of a brief. Once it is there, ask Director to add more reference points to it, and the location comes back with four images on it instead of one. You never open a panel, and the location Element it hands back is ready to reference in the next thing you generate.

By hand. Click the plus icon in the toolbar and pick the Location Scout app. Describe your location, attach the reference you already have on the canvas, toggle on the four additional views, and hit Generate.

The manual route is worth knowing for the same reason it always is. When a location really matters and you want to run it yourself, the controls are right there. For everything else, asking Director is faster and it is the route to start with.

What happens in the background

This is the part worth understanding, because it explains why the four views match rather than merely resembling each other.

Location Scout does not generate four images. It generates one camera move.

From your starting image, MITO builds a five-second shot that travels around the space, and takes a single frame at each point where the camera settles. The four stops land roughly seventy degrees apart, so together they cover the room. Each frame is then restored to full detail and colour-matched back to the original.

The reason this works is what the camera does not do. The angle changes, and the height and the distance stay exactly where they were. The camera is circling one room rather than being teleported to four new ones, and because every view is a frame of the same continuous shot, the layout and the lighting carry across them on their own.

Four separate generations could never promise that. They would each start over.

What it costs

A set is 545 credits, and it takes about two and a half minutes end to end. Prices as of October 2026.

Worth putting that next to the alternative. Generating four location references by hand means four prompts, several re-rolls each when they come back as different rooms, and a pass at the end to pick the ones that nearly agree. The credits are the smaller half of what that costs you.

Locations and characters are the same problem

If you have read how to keep AI video characters consistent, this will feel familiar, because it is the same idea pointed at a room instead of a person.

A character needs to be the same character in shot four. A location needs to be the same location. Both break for the same reason, which is that one image leaves too much unsaid, and both are fixed the same way, by giving the model more of the thing to look at before it starts guessing.

Elements is where both of those live. Location Scout is what makes a location Element worth referencing.

The short version. One image of a room tells the model about one wall. Location Scout walks a camera around it and brings back four views, so the next shot you generate is in the place you already built.

It is exclusive to MITO, and it is live now.

Open a project and scout a location →