EEG · Seizure detection · Decision support
Every second of EEG, read.
Neurazi screens EDF recordings with a deep-learning model and flags likely epileptic seizure activity — so your clinicians review the right recordings first.
DECISION SUPPORT — NOT A DIAGNOSIS
recording_04.edf · patient P00124 · 182 windows
Seizure DetectedConfidence
96.40%
Peak prob.
96.4%
Mean prob.
31.2%
01 — Protocol
How a recording becomes an answer
01
Upload the recording
A standard EDF file from any EEG system, tied to an anonymous patient record in your hospital's workspace.
02
The model reads every window
A deep-learning model screens the recording in ten-second windows and scores each one for seizure activity.
03
Review with full context
Classification, confidence, and per-window probabilities — alongside the patient's complete analysis history.
02 — Platform
Built like clinical software, because it is
Confidence, not just labels
Each analysis reports peak and mean seizure probability across every ten-second window — so clinicians can judge borderline cases instead of trusting a bare label.
0s
analysis windows — every second of the recording is scored
0
patient names stored — records are anonymous by design
Isolated workspaces
Every hospital sees only its own patients and results — isolation enforced at the API level.
Audit trail
Access, sign-ins, and exports are logged — who, what, and when — for accountability.
Timelines & reports
Every analysis builds the patient's history — exportable as PDF or Word for referrals.
See Neurazi on your own recordings
We'll set up a workspace for your team and walk through an analysis together — with your EEG files, not a canned demo.