Visko launches Orbis for continuously generated interactive video

Visko has launched **Orbis**, a foundation model built around a different interaction pattern from conventional text-to-video systems: instead of generating a finished clip and stopping, Orbis is designed to keep a generated world running while the user changes what happens next.\n\nThe company calls this a “Live Model.” On Visko’s current product surface, Orbis supports text-to-world and image-to-world generation, interactive story changes, voice-driven generation, live streaming and simulation-oriented demos. The practical distinction is steerability during generation: users are meant to intervene without restarting a new isolated clip.\n\nThat is a meaningful product idea, but the launch claims need to be separated from independently demonstrated performance. Visko says Orbis can stream at 4K and 24 frames per second, react immediately to interventions and sustain hour-scale generation without obvious visual or color drift. SiliconANGLE independently confirms the launch, funding and product positioning, but those performance figures originate with Visko and should be treated as vendor-reported until reproducible testing or broader independent evaluation exists.\n\n## Public access still has an account gate\n\nVisko’s launch release says it opened public access to Orbis. The current product site does expose multiple Orbis demos and a freeform playground entry point, but selecting the freeform experience redirects to a login screen. That makes the launch more accessible than a private research preview, while still falling short of anonymous, frictionless hands-on access. Buyers and researchers should distinguish “publicly available to sign up” from “open demo with no account gate.”\n\n## Why the continuous model matters\n\nMost generative-video products are optimized around discrete outputs: prompt, render, inspect, then regenerate or edit. Orbis is trying to make the generated environment itself persistent. If that approach holds up under real use, it could matter beyond entertainment video, including game prototyping, robotics simulation, interactive training and live creative tools where state needs to survive across user interventions.\n\nVisko also announced a **$10 million pre-seed round led by Llama Ventures** alongside the launch. Funding is useful context for the company’s runway, but it does not validate the model’s technical claims. The central question is whether Orbis can preserve physical, visual and narrative state reliably across long sessions under diverse prompts, not whether the launch demo looks impressive.\n\nFor now, Orbis is worth watching because it pushes generative video toward a continuously running system rather than a sequence of isolated clips. AiToolMap would still want independent evidence on long-session coherence, latency, output resolution under load, pricing and practical access limits before treating the strongest launch claims as established product performance.