Roadmap
What is planned, what is being weighed, and what VT Atlas will not do. The last list is the useful one: knowing where a plugin stops is worth more than a wish list.
The dates are targets, not commitments. They exist because "sometime" is not useful when you are deciding whether to build on this, and a quarter is honest about the precision actually available. An item slips rather than ships half finished, and when one slips this page says so.
Shipping now — v1.0, Q3 2026
Everything the plugin does today. It is on the timeline for the same reason the later stops are: so you can see at a glance what you get now against what you would be waiting for.
Regions from your gameplay tags, four volume shapes including splines, per-player replicated discovery, region features, per-volume user data, and exits that pair with enters however an actor left. See Concepts for what each of those means.
Planned
v1.1 — Q4 2026
Save and restore discovery. Discovery is per player and lives in memory today, so a map that filled in during a session starts empty in the next one. The plan is a pair of functions that hand you the discovered set as plain data and take it back, so it goes in whatever save system you already have rather than one this plugin invents.
An editor region browser. A panel listing every region in the level with its tag, volume count, depth and current occupants, so you can find the one you mean without hunting the outliner. Today the gameplay debugger shows this at runtime and the editor shows nothing.
v1.2 — Q1 2027
Region volume presets. Placing a volume still means setting a tag, a shape and a tracked-object list every time. A preset asset would carry a configured starting point, which matters most for teams where one person defines the conventions and everyone else follows them.
A spatial index for very large levels. The work is already done and measured: at a hundred regions a plain scan answers in a fifth of a microsecond and an index would be slower overall once streaming cost is counted. Past a few thousand regions that reverses. A uniform grid is the measured winner and it ships behind a setting when somebody has a level that needs it. See Performance for the numbers behind that decision.
This one sits late on purpose. Nothing in it helps a level that does not need it, and the measurement says most levels do not.
Considering — 2027 onward, no version
Real ideas with real arguments against them. They carry no date because they ship only if the argument for wins, and some of them will not.
A Gameplay Ability System bridge. Granting an ability or applying an effect while an actor is in a region is an obvious fit, and it was deliberately left out of v1: VT Atlas has no GAS dependency today and adding one makes every buyer carry it. The likely answer is a separate optional module rather than a dependency in the core plugin.
Region-scoped audio helpers. Crossfading ambience on region edges is one of the most common things people build with this. It is also five nodes, and a helper that saves five nodes but only fits one project's audio setup is worse than the five nodes.
Volume shapes from static meshes. Spline regions already cover outlines that a box cannot. Taking the shape from an arbitrary mesh is possible and would mostly be used to make regions far more expensive than they need to be.
Per-region net relevancy hints. Regions know who is where, which is exactly what relevancy wants. This one is genuinely hard to do without making promises about netcode that the plugin cannot keep.
Not planned
Deliberate limits, not gaps. No dates, because these are not coming.
A region vocabulary. VT Atlas will never ship a set of region names or a depth convention. Regions are your gameplay tags and the hierarchy is yours; a built-in vocabulary would be a second place to say the same thing.
Replicated region state. Membership is computed identically on every machine from level content everyone loads, so there is nothing worth sending. Discovery replicates because it is genuinely per-player knowledge. Nothing else will.
Pathfinding, navigation or AI behaviour. Regions can tell an AI where it is. What it does about that belongs in your behaviour tree, not here.
A quest or objective system. Regions are a good trigger for objectives and a bad place to keep them.
Asking for something
If you need one of these sooner, or something that is not here, say so through the contact form. Knowing that a real project is blocked on an item moves it up this list faster than anything else, and it is the only thing that reliably does.