Simulon’s Characters Now React to Your Real Footage

Most AI video tools start with a prompt. Simulon starts with your phone camera. On Oct 28, 2026 the iPhone app shipped Reality-Aware Rendering: you place a 3D character in a real scene, act out the shot, and when you stop recording Simulon renders the take with the character reacting to what you did, what you said and the place you filmed.

The launch clip from founder Divesh Naidoo passed 1.5 million views on X in two days. Here is what the feature actually does, what it costs, and why it matters for anyone building creative AI products.

What Reality-Aware Rendering does

Naidoo's launch post describes the flow in three steps:

• Place. You put a neural asset, Simulon's term for a 3D subject it generates, into the scene with your phone in AR. Its placement sets the starting position and scale.
• Perform. You record the scene. During recording the asset turns into a "ghost", so you direct it with actions, gestures and spoken instructions instead of a typed prompt.
• Render. When you stop recording, Simulon reads the whole take and generates the final shot. Naidoo stresses that the result is regenerated: "a new performance of the scene, not a copy of the footage."

In an earlier demo, Naidoo placed a raptor next to a pond and dropped a stone into the water to see how it would react. Simulon, he wrote, "uses the recorded scene to generate his behaviour and integrate him into the footage afterward."

Zero prompting, real footage

The launch short was, per Naidoo, "made with zero prompting. Just real footage, 3D neural assets and reality-aware rendering." His side-by-side cut shows the Simulon render on top and the raw phone recording underneath, where the character simply isn't there.

Assets don't have to come from a 3D artist. A day after launch, Naidoo turned a single image of Dracula into a neural asset and had it ride a real RC car across his dining table; when the car went over the edge, the character jumped off and stayed on the table. "All of this happens procedurally, with no prompting," he wrote.

What creators are noticing: light and shadow

The hard part of putting a CG character into phone footage has always been matching the light. That is what early testers keep pointing at. VFX artist and filmmaker Aharon Rabinowitz posted a test where the character's shadows matched the room's existing ones.

These are short demos picked by the founder and fans, so treat them as best cases. We have not run our own tests, and Simulon has not published failure rates or limits for how long a take can be.

Built on Simulon 2.0

Reality-Aware Rendering sits on top of Simulon 2.0, which arrived on the App Store on Oct 9, 2026. According to Simulon's site and its App Store listing, 2.0 lets you:

• Clone reality: turn real photos, such as a pet, a car or a sculpture, into 3D neural assets.
• Go from sketch to real: turn a drawing or a few words into an asset.
• Combine assets into something new, and keep them in a personal library.
• Capture and animate: find your angle in AR, capture a photorealistic image, then describe an action, edit the scene, or set start and end frames to make a video. Processing happens in the cloud.

The App Store lists version 2.1.1, released Oct 28, 2026, for iOS 17 or later, from Beamm Technologies Inc. Android is "coming next", with no date given.

Pricing

The app is free to download. As listed on simulon.com on Oct 30, 2026:

• Free: 100 one-time starter credits to try subjects from the Simulon Library. Creating your own assets needs a paid plan.
• Simulon Plus, $14.99/month: custom subjects from words or images, monthly usage, commercial use for individuals.
• Simulon Pro, $39.99/month: 3x Plus usage, studio-tier video generation, up to 15% off top-ups, commercial use for businesses.
• Top-ups: 100 credits for $4.99, 260 for $12.99, 500 for $24.99, for Plus and Pro subscribers. Credits don't expire.

Simulon doesn't say how many credits a Reality-Aware render uses, so check the in-app costs before you plan a project around it.

Why it matters for builders

Most generative video products compete on the same input, a text box. Simulon's bet is that the camera is a better interface: timing, framing and spatial context come from the recording, and the model fills in the character. That is a product decision as much as a model one, and it points at a few things worth watching:

• Input beats prompt. Performance capture from a phone removes the "slot machine" loop of re-prompting, which is where many video tools burn users' credits and patience.
• Mobile-first creation. Shipping on iPhone first, with cloud rendering, keeps the hardware bar low. Android support will decide how big the audience gets.
• Subscription plus credits. Plus/Pro tiers with non-expiring top-ups mirror how other AI creative apps now price heavy GPU work. Compare it with the plans in our AI tools comparison for 2026, or with how Figma started billing for its agent.

Verdict

Reality-Aware Rendering is one of the more original takes on AI video this year: direct a scene by acting it out rather than describing it. The demos are striking, especially on lighting, but they are curated. If you have an iPhone, the free credits are enough to judge it on your own footage; custom assets and heavier use need Plus at $14.99 a month.

Sources

• Simulon: official site and pricing
• Simulon on the App Store
• Simulon on X: Reality-Aware Rendering launch (Oct 28, 2026)
• Divesh Naidoo on X: Introducing Reality-Aware Rendering (Oct 28, 2026)
• Divesh Naidoo on X: how it works (Oct 20, 2026)
• Divesh Naidoo on X: zero-prompting side by side (Oct 28, 2026)
• Divesh Naidoo on X: Dracula from a single image (Oct 29, 2026)
• Aharon Rabinowitz on X: shadows test (Oct 30, 2026)
• Tim Soret on X (Oct 20, 2026)
• Divesh Naidoo on X: Simulon 2.0 on the App Store (Oct 9, 2026)