On 2026-11-05 World Labs released research preview Frameboy: a highly optimized Atlas variant streamed from cloud GPUs to the browser at up to ~15 fps (vendor claim) with live camera control. A public demo and a Viewfinder closed-beta overlay shipped together; CEO Fei-Fei Li amplified “Real time Atlas” on X.

Key Takeaways

  • ✓Official: World Labs shipped Frameboy research preview on 2026-11-05 (blog + X)
  • ✓Vendor claim: optimized Atlas streams from cloud GPUs at up to ~15 fps with live camera control
  • ✓Public demo: frameboy.worldlabs.ai; Viewfinder closed beta builds scenes then overlays live explore
  • ✓Capabilities: single-image/text scene seeding + multiview novel-view synthesis; RTFM form factor at Atlas scale
  • ✓Vendor notes powerful cloud GPUs required and cost vs local rendering; no independent quality/latency board
World Labs Ships Frameboy: Optimized Atlas Streams at up to ~15 fps for Real-Time Browser World Exploration
🖼️Official Media
Click to view high-res
🧭

Turn your technical choice into a development budget

Compare 40 dev plans & simulate token costs vs $20/mo subscriptions

🔬

In-Depth Technical Analysis

Core Background & Industry Pain Points

If each world-model render takes tens of seconds or minutes, users click cautiously. On 2026-11-05 World Labs (co-founded by Fei-Fei Li) shipped research preview Frameboy: an optimized Atlas variant streamed from cloud GPUs to the browser at up to ~15 fps (vendor claim) with live camera control, also wired into closed-beta Viewfinder.

Architecture Highlights & Internals

Atlas is an autoregressive world model with pixel-perfect camera control, single-image/text seeding and multiview novel-view synthesis. Frameboy serves a fastest-inference Atlas variant; the public demo offers curated scenes, while Viewfinder builds scenes frame-by-frame (with depth) then overlays live explore. It remakes last year’s RTFM form factor at Atlas scale.

Authoritative Benchmarks & Measured Scores

No independent quality/latency board; the headline figure is cloud streaming “up to ~15 fps” (vendor claim). The blog argues slow generation forces cautious planning while real-time interaction drives trial cost near zero. World Labs notes powerful cloud GPUs are required and cost is “not cheap” versus local rendering. Do not invent scores.

Developer Hands-on Guide

Demo: frameboy.worldlabs.ai. Custom scenes need the Viewfinder waitlist. Blog: worldlabs.ai/blog/atlas-realtime. Useful for film scouting, level concepts and landmark walkthroughs. Caveat: fps is a vendor upper bound; measure control-to-photon latency separately; cloud cost remains high.

Action HubReady to adopt this in production?

Benchmark side-by-side against alternatives, or calculate monthly token cost vs subscription break-even.

Up NextWorld Labs launches Atlas, an omni world model for spatial intelligence
Read Now→