NeuroStudio
Local training and experiment history inside the Suite GUI. Configure runs, compare metrics, and export portable models without a cloud training requirement.
Overview
NeuroStudio is where training and export live inside NeuroTendrils Suite. After NeuroFlow produces clean artifacts, the next friction is usually the same: a thin script, a forgotten hyperparameter, a metrics folder that only one person understands, and another afternoon spent reconstructing what last week's run actually did.
In NeuroStudio you configure a run, watch metrics as it progresses, browse past experiments, and export models in portable formats such as ONNX. Training is designed to run on the user's machine (CPU, GPU, or Apple Silicon). A cloud training account is not required for core use, so iteration stays close to the data and the rest of your local workflow.
From cleaned data to a portable model
Walk the training loop. Config, train, and export each look different on purpose.
Start from artifacts
Bring in cleaned EEG from NeuroFlow or compatible local exports so training begins from a known preprocessing state.
What you can do
Select a capability to see how training stays local and comparable.
Configure, run, and compare training loops in one place instead of jumping between scripts, terminals, and ad hoc notes.
Who it fits
Built for people who need comparable local experiments, not a hosted ML platform.
Iterate on models next to the data and keep experiment history readable for the next collaborator.
Part of NeuroTendrils Suite
NeuroTendrils Suite is research and education software in active development. It is not a medical device, not clinical software, and does not make clinical claims. Core use is designed to run locally. NeuroFlow is open source; other Suite modules are not.