Neural data infrastructure for people who build things
MachineDigit builds the layer between the brain and software: wearable, non-invasive EEG hardware, the signal processing that turns raw microvolts into usable states, and the APIs developers plug into. Not a gadget, not a medical device, and not a promise of mind reading. An honest platform for working with brain-signal data.
We turn EEG brain-signal data into software inputs
The system has three parts that work as one. Dry-contact EEG wearables capture brain-signal data with no gel and no skin preparation. A signal-processing engine filters and decodes that data into real-time states: a focus and attention index, EEG band power, and signal quality. WebSocket and REST APIs, with Python and JavaScript SDKs and raw EEG export, expose all of it to whatever you are building.
Everything is labeled by maturity. Specs are published as they are validated, reliable detections are separated from experimental ones, and session recording happens only with per-session consent. Brain-signal data is treated as sensitive data everywhere we operate, is never sold, and is never used for advertising.
Brain-computer interfaces are stuck between the lab and the hype
Research-grade EEG means gel, caps, and clinical setups that nobody wears for a full working session. Consumer neurotech overpromises and locks its data away. Builders who want to experiment with brain-signal input are left with hardware they cannot access programmatically or claims they cannot trust. MachineDigit exists for exactly that gap: wearable hardware people actually keep on, honest published specs, and open access to the data and the decoded states.
Building experimental interfaces, games, and neuroadaptive applications against live APIs.
In BCI, neuroscience, HCI, and XR, with raw EEG export in EDF, CSV, or JSON.
Exploring hands-free, signal-driven input for people the current input layer leaves out.
Evaluating neuroadaptive features with a platform that states plainly what works today.
Two wearables, one platform
The temple patch. Four dry-electrode channels at 256 Hz, worn above the eyebrow, under 15 g. The higher-fidelity option for sessions that need more signal.
The ear-worn unit. Two dry-electrode channels at 256 Hz, worn at the ear, under 20 g. The discreet option for everyday, all-session wear.
Signal engine, WebSocket and REST APIs, Python and JavaScript SDKs, raw EEG export, and the beta dashboard for running sessions and experiments.
In beta, and honest about it
The hardware is Rev A and the beta program is open and selective: developers, researchers, accessibility partners, and early adopters, admitted individually. Some detections are reliable, others are experimental, and the spec sheet marks every field validated or in validation. That labeling discipline is the product philosophy in miniature. If we have not measured it, we do not claim it.
Who is behind MachineDigit
Monika founded MachineDigit to make brain-signal data something builders can actually work with. Before MachineDigit she spent over a decade in C-level leadership at Cymonz, an international payments platform, where the same discipline this site runs on was formed: regulated data, honest claims, and infrastructure other companies build their products on.