NimbusNimbus
Back to Blog

May at Nimbus Studio: EEG Devices and Interactive Games

May was all about making Nimbus Studio easier to organize, easier to collaborate in, and much more fun to test in real time — with major upgrades to the Home screen, desktop app, live hardware streaming, and Interactive Games.

What’s new (May 2026)

  • Live EEG streaming for BrainBit (full family) and PiEEG (plus live EEG calibration improvements)
  • Interactive Games for real-time feedback when training and testing classifiers
  • A redesigned Home that organizes work into Projects and Templates
  • Lab sharing so teams can collaborate and review results together
  • A faster, more reliable Desktop app with local runs that sync to the cloud (building on the February desktop app release)
  • More built-in models and improved evaluation summaries

Live EEG: BrainBit and PiEEG streaming + better calibration flow

May’s release brings a big leap for real-time work: improved hardware support and clearer “are we ready to record?” feedback.

BrainBit (via NeuroSDK)

  • Full BrainBit family support (desktop + deploy)
  • Device picker before connecting
  • Live telemetry (battery, signal strength, motion, saturation, dropped samples)
  • Auto-reconnect with visible gaps on the EEG strip when the link drops
  • Gain profiles (low / medium / high) for flat or hot signals
  • Optional hardware markers (TTL triggers)
  • More quality checks (bridged-electrode detection and reference-channel warnings)

PiEEG

  • Stream from PiEEG in the desktop app with a live calibration preview

Calibration improvements

  • Trial protocol presets with estimated session duration (see BCI calibration: from hardware to trained decoder)
  • A clearer live EEG preview and readiness feedback before recording
  • More explicit signal quality indicators (mains hum, reconnect gaps, reference-loss alerts)
  • Trigger presses now register every press (not only the first)

Interactive Games: real-time feedback that makes training and demos easier

The new Interactive Games node turns classifier output into something you can see and control immediately. When your pipeline includes it, Deploy opens a game-first fullscreen experience.

Game presets included

  • MI Arrow Controller (2–4 classes)
  • MI Target Hit (adjustable speed)
  • MI Spoon Bend (3D demo)
  • MI Space Invaders (left/right MI)
  • MI Doom (desktop) (Doom arena with guided training)

Connect to external apps (optional)

If you want to drive something outside Nimbus Studio, you can send decisions over OSC or LSL to tools like VRChat, TouchDesigner, LabRecorder, and more.

Tip: Place Interactive Games after your model and decision policy, before results output.

Home: projects, templates, and sharing (so your work doesn’t turn into a long list)

If you’ve been building multiple pipelines (different paradigms, studies, or experiments), Home now gives you structure from the moment you sign in.

Projects

  • Group pipelines into Projects instead of a flat “everything” list
  • Quickly jump into the canvas, duplicate pipelines, and browse your recents
  • Save / load with an inline project picker (switch or create projects without leaving what you’re doing)
  • Start from Templates: curated starters plus a full template catalog

Lab sharing & collaboration

Sharing is now built into the workflow — not an afterthought.

  • Invite collaborators by email with roles: Owner, Editor, Viewer
  • A Shared with me section on the sidebar shows everything you’ve been invited to
  • Shared Executions: teammates can watch live progress, view summaries/results, and review generated artifacts
  • Viewer read-only canvas: viewers can inspect configurations safely without editing
  • Duplicate shared work instantly (pipelines or entire projects) to build on a colleague’s setup

More Deep Learning models + clearer evaluation summaries

To help you benchmark and iterate faster, Nimbus Studio now includes more model options and improved evaluation views:

  • New/expanded model nodes including NeuroRVQ, EEGNet variants, EEG Conformer, EEGNeX, ATCNet, and TRCA (SSVEP)
  • Stronger cross-validation and multi-subject benchmarking for clearer aggregate comparisons
  • Clearer band-power charts in spectral feature nodes

(If you’re not sure which model to start with, templates are the quickest path — then iterate from there.)

Desktop app: run locally, stay in sync

The desktop app now feels much closer to a “native” development workflow:

  • Sign in via your browser (GitHub, Google, or email), then the app reopens automatically
  • Run locally, sync to the cloud when you’re back online (with clear sync status and a retry option)
  • Auto-update progress so you can see downloads and installs happening

If you’re using hardware streaming or Interactive Games, the desktop app is the smoothest way to work.

Plans + everyday polish

  • Free tier: limited monthly runs (deploy streaming UI still available)
  • Pro: unlocks calibration and full runtime access (you’ll see a pricing page if you hit limits)

Plus a long list of “small things that matter”:

  • A smoother canvas with a pinned header + terminal
  • Better mobile/tablet experiences on hub and project pages
  • Richer node visualization tabs

Getting started