NeuroSkill — Local-First EXG/BCI Neurofeedback Workbench

A local-first desktop app that turns real-time EXG streams into GPU band-power analysis, neural embeddings and session-level BCI research views.

Stars
★ 107
Last updated
4d ago
License
GPL-3.0
Primary language
Rust

At a glance

How it runs
Desktop appCLI
Works with
Portable with changes
Cost
Free, no paid service needed
Setup effort
Medium · a few setup steps
You'll need
Hugging Face CLI (hf)npmRust/Tauri v2 toolchainSvelteKitTauri v2Shell / CLINetwork accessLocal filesystem
Typical use
A neurofeedback researcher who wants live multi-channel band power and session comparisons on a laptop without uploading raw EEG to a cloud service.
Not a fit if
  • Teams needing an FDA/CE-approved medical device for diagnosis or treatment
  • Users unwilling to set up the Hugging Face CLI, Rust/Tauri and npm toolchains
  • Users who want a pure cloud SaaS rather than local model weights and EXG devices
Website
neuroskill.com ↗ set on GitHub
Maintenance
Commits in the last 30 dayslast commit 2026-11-17latest release v0.1.2 (2026-11-13)

What does this agent do, and when should you use it?

NeuroSkill is an open-source, local-first desktop neurofeedback and BCI application built with Tauri v2 (Rust) and SvelteKit, explicitly scoped to research use and not a medical device. It ingests real-time EXG streams from supported headsets or any compatible LSL source, as well as remote rlsl-iroh streams, and lists devices from BrainBit, OpenBCI, Emotiv, InteraXon Muse, g.tec, Neurosity, NeuroSky, Ōura and more. On top of the live signal it performs GPU band-power analysis and neural embeddings, then supports session comparison, sleep staging, UMAP visualization, labeling, similarity search and screenshot-based search. A catalog defined in src-tauri/exg_catalog.json drives loading of EEG/EXG foundation models such as ZUNA, LUNA, REVE, LaBraM, EEGPT, CBraMod and SleepFM from Hugging Face. The app also embeds a local LLM and TTS and exposes a WebSocket/HTTP API, and ships as macOS dmg, Windows exe and Linux AppImage releases or via a Homebrew cask. Because everything runs in the desktop process, raw physiological data does not have to leave the machine.

At runtime NeuroSkill reads live multi-channel EXG data from a supported device or from any LSL source, including remote rlsl-iroh streaming, and renders it in a real-time desktop view. It then runs GPU band-power analysis and neural embedding computation on captured signals. Recordings become sessions that can be compared against each other, staged for sleep, projected with UMAP, labeled, and retrieved through similarity search or screenshot-based search. The model layer is declared in src-tauri/exg_catalog.json and pulls pretrained weights from Hugging Face (ZUNA, LUNA Base/Large/Huge, REVE Base/Large, OpenTSLM, SensorLM, SleepFM, SleepLM, OSF Base, SignalJEPA, LaBraM, EEGPT, TRIBE v2, NeuroRVQ, CBraMod, ST-EEGFormer Small/Base/Large) for downstream inference. A local LLM plus TTS integration and a WebSocket/HTTP API make the app programmable from outside. Accompanying docs cover architecture, device integration, metric formulas, hooks, LLM internals, LSL integration, the API surface, and platform build steps.

  1. A neurofeedback researcher who wants live multi-channel band power and session comparisons on a laptop without uploading raw EEG to a cloud service.
  2. An ML engineer benchmarking several EEG foundation models (LaBraM, EEGPT, CBraMod, LUNA) who needs one acquisition pipeline that feeds loading, embedding and visualization.
  3. A sleep or cognition lab using SleepFM/SleepLM-style models with overnight recordings for sleep staging and UMAP inspection.
  4. A hardware prototyper working with Muse, OpenBCI Cyton, Emotiv or g.tec Unicorn who needs a desktop app with broad device support and a WebSocket/HTTP API.
  5. A team wiring a local LLM and TTS into a physiological-signal workflow through the app's hook and local-model pipeline.
  6. An instructor or demo presenter who needs a one-command install on macOS, Windows or Linux and a live BCI walkthrough.

Key setup facts

Pulled from the install and usage notes below. The project README has the current values.

Install / run
  • curl -LsSf https://hf.co/cli/install.sh | bash
Config files
src-tauri/exg_catalog.json
Install notes cover
Windows · macOS · Linux

How do you install or deploy this agent?

Prebuilt releases are published for each desktop platform; download and launch.

# macOS
https://github.com/NeuroSkill-com/skill/releases/latest/download/NeuroSkill.dmg
# Windows
https://github.com/NeuroSkill-com/skill/releases/latest/download/NeuroSkill.exe
# Linux
https://github.com/NeuroSkill-com/skill/releases/latest/download/NeuroSkill.AppImage

On macOS you can also install via Homebrew:

brew tap NeuroSkill-com/skill && brew install --cask neuroskill

Building from source starts by installing the Hugging Face CLI and downloading a model (ZUNA shown):

curl -LsSf https://hf.co/cli/install.sh | bash
hf download Zyphra/ZUNA

On Windows the Hugging Face CLI install command is:

powershell -ExecutionPolicy ByPass -c "irm https://hf.co/install.ps1 | iex"

Then run the repository setup script and start the dev app:

npm run setup -- --yes
npm run tauri dev

And build a production bundle:

npm run tauri build

How do you use this agent?

Launch the desktop app, pick the EXG device or LSL stream you want (BrainFlow, MATLAB and pylsl sources all qualify via LSL), and start the real-time stream view. Save the recording as a session, then use the built-in tools to compare sessions, run sleep staging, generate UMAP visualizations, apply labels, and run similarity or screenshot-based search over the data. To run a specific foundation model, make sure its weights were fetched with hf download first; the authoritative list lives in src-tauri/exg_catalog.json. External scripts talk to the app through the WebSocket/HTTP API documented in docs/API.md, while local LLM and TTS wiring is covered in docs/LLM.md and extension points in docs/HOOKS.md. The device and model lists in the README are auto-generated and should be refreshed with npm run sync:readme:supported after changing support.

What are this agent's strengths and limitations?

Pros
  • Local-first by design: acquisition, model inference and storage happen on the desktop, which suits researchers handling sensitive raw physiological data.
  • Broad hardware coverage: the README lists BrainBit, OpenBCI, Emotiv, Muse, g.tec, Neurosity, NeuroSky, Ōura and others, plus any compatible LSL source.
  • Auditable model catalog: src-tauri/exg_catalog.json enumerates 20 EEG/EXG foundation checkpoints across ZUNA, LUNA, REVE, LaBraM, EEGPT, CBraMod, SleepFM and more.
  • Programmable surface: a WebSocket/HTTP API, local LLM and TTS integration, and a documented hooks mechanism make it embeddable in existing pipelines.
  • Complete three-platform distribution: macOS dmg, Windows exe, Linux AppImage and a Homebrew cask keep install friction low.
Limitations
  • Explicitly not a medical device: no FDA/CE approval means it cannot be used for diagnosis or treatment, which blocks clinical adoption.
  • You must supply your own EXG hardware and Hugging Face weights, and source installs require the Hugging Face CLI, npm and the Tauri/Rust toolchain.
  • GPL-3.0-only licensing constrains closed-source redistribution or embedding in commercial products.
  • The repository documents no server or container deployment path, so it cannot run as a headless service on a cluster.
  • The README does not state which LLM/TTS models are used, their hardware acceleration requirements or quantization options, so inference performance cannot be estimated from the docs.

How does this agent compare with similar options?

Key facts side by side with the most closely related agents.

Agent Source review Form / cost Stars Updated Language Full support on
NeuroSkill — Local-First EXG/BCI Neurofeedback Workbench This agent 57 · Major gaps Desktop appFree ★ 107 4d ago Rust —
Open Science Workbench 81 · Good Desktop appFree + model costs ★ 5.6k 1d ago TypeScript Codex · Claude Code
Bough Work and Cost Explorer 78 · Good CLIFree ★ 133 2d ago Go Codex · Claude Code
ChatLab — Chat History Analyzer 70 · Some gaps Desktop appFree + model costs ★ 7.7k 2d ago TypeScript —

How does FollowAgents rate this agent?

FollowAgents source review · FARS-2.1
Major gaps
57/ 100 5-point scale 2.9 / 5
This review covers the source as of Nov 08, 2026. The project has shipped 1 stable release(s) since, which have not been re-reviewed. The score and verdict describe the reviewed version only. Recent releases ↓
Trust 13/29
Reliability 5/14
Adaptability 12/18
Convention 14/18
Effectiveness 9/13
Verifiability 4/8
Why each dimension lost points
Trust13 / 29 · 2.2/5

The README describes a local-first EXG/BCI desktop app and clearly limits it to research, not medical use. SECURITY.md documents TOTP pairing for remote access, read scope by default, client revocation, and the unencrypted credential file's location and permissions. Deductions: the README also lists remote tunnels, HTTP/WebSocket APIs, cloud-streamed devices, and external model downloads, but this sample does not give a complete data-flow or permission-boundary account; failed pairing can trigger a global lockout, with recovery dependent on daemon restart; publisher identity and maintenance responsibility are not verified.

Reliability5 / 14 · 1.8/5

Cargo.toml and the CI workflow describe dependency unification, patches, and build checks, showing concrete reliability-oriented maintenance. Deductions: Cargo.toml relies on multiple Git branches and local patches, and CI refreshes the rlx main branch, limiting dependency availability and reproducibility; the README, SKILL.md, and sampled code do not cross-check all product behavior, and the sample provides little evidence about implemented failure messages.

Adaptability12 / 18 · 3.3/5

The README and SKILL.md cover several EXG/BCI devices, desktop systems, connection methods, and research, sleep, comparison, and automation scenarios; they distinguish fNIRS from traditional EEG. Deductions: support for some devices depends on bridges, feature flags, or cloud streaming, and the sample does not systematically explain those boundaries; the device-agnostic claim coexists with device-specific differences, leaving triggers and hardware requirements incompletely specified.

Convention14 / 18 · 3.9/5

The README provides a structured documentation index, development quickstart, and setup requirements; SKILL.md has a contents list, command reference, and use-case recipes. LICENSE and package metadata identify GPL-3.0-only; the README points to versioning and changelog files, and CI includes a version synchronization gate. Deductions: the sample does not identify a specific maintainer or team responsibility; some setup relies on external CLIs, model repositories, and platform tools; the product and CLI names are not fully consistent.

Effectiveness9 / 13 · 3.5/5

SKILL.md offers human-readable summaries, pipe-safe JSON, and full output modes, with status polling, queries, comparisons, sleep analysis, and automation recipes. These appear useful to users with a running local daemon. Deductions: the tool requires compatible devices, a desktop app and daemon, and potentially model or GPU resources, making setup and operation costly; the sample shows broad functionality but cannot establish its incremental user value.

Verifiability4 / 8 · 2.5/5

The README ties generated device and model lists to generation markers and a canonical model catalog, and ties product versioning to VERSION; SECURITY.md gives specific, checkable claims about TOTP parameters, replay prevention, lockouts, and secret storage. Deductions: the sampled files omit the referenced canonical data, implementations, and complete API documentation; the included tests cover only some frontend types and calibration logic, leaving several security and feature claims supported by a single source in this packet.

Risks and how to mitigate them
  • The product is explicitly research-use only and is not a medical device; do not use its outputs for diagnosis or treatment.
  • EEG, PPG, activity, and screenshot data may be sensitive. The security document says TOTP credentials are unencrypted, and granting a remote client full scope exposes the entire local API.
  • Failed remote pairing attempts share a lockout counter across unregistered clients and can delay legitimate pairing; device support and model execution also depend on hardware, bridges, and local resources.
Evidence confidence: Low Reviewed Nov 08, 2026 Reviewed revision b5be79d3bf08 New commits since this review; the score may not cover them
See the full review method →
Maintain this project? Add the FARS badge to your README

The badge shows the current review score and verdict and links back to this page. It refreshes about once an hour after a re-review.

FARS

[![FARS review](https://followagents.com/badge/neuroskill-com-skill.svg)](https://followagents.com/en/agents/neuroskill-com-skill)
<a href="https://followagents.com/en/agents/neuroskill-com-skill"><img src="https://followagents.com/badge/neuroskill-com-skill.svg" alt="FARS review"></a>

Recent releases

Latest entries from the project's GitHub Releases. Notes are quoted from upstream as written, not rewritten or tested by FollowAgents.

  1. v0.1.2 Skill v0.1.2
    Release notes excerpt
    Changelog Features - EmotiBit support: connect an EmotiBit board over Bluetooth (emotibit-ble firmware) or WiFi (stock firmware). Heart rate, inter-beat intervals, SpO2 and skin temperature record as device vitals, accelerometer and gyroscope as motion, and battery as battery. A BLE board is found by its GATT service; add a WiFi board as emotibit:<id or ip>. …
    View on GitHub ↗
  2. v0.1.1 Skill v0.1.1
    Release notes excerpt
    Changelog Bugfixes - A pairing code now works only once. Registering a remote device over iroh used to accept any TOTP code that was still within its ±30 s window, so a code someone had seen or intercepted could be reused for about a minute. Each TOTP credential now remembers the time step of the last code it accepted and rejects that step or any earlier one with "otp already used; wait for the next code" (RFC 6238 §5.2). Pairing a second device, or re-registering one, with the same credential therefore waits for the next code, up to 30 s. See [SECURITY.md](SECURITY.md). …
    View on GitHub ↗
  3. v0.1.1-rc.4 Skill v0.1.1-rc.4
    Release notes excerpt
    Changelog Features - Empatica EmbracePlus: connect the wristband over BLE via the empatica crate. Skin temperature streams as vitals; EDA, steps and ambient light are logged as metadata. Compressed PPG/IMU are not decoded live. Contributors - Eugene Hauptmann
    View on GitHub ↗

FAQ

Does it cost anything?
No. NeuroSkill is free and open source under GPL-3.0-only, with release downloads and a Homebrew cask. Your real costs are EXG hardware, local compute, and the network/storage needed to pull model weights from Hugging Face.
Can I use it in a clinical setting?
No. The README states plainly that it is for research use only and is not a medical device, with no FDA/CE approval and no permitted diagnostic or treatment use.
Will my existing EEG headset work?
Check the supported device list first (BrainBit, OpenBCI Cyton/Ganglion, Emotiv, Muse, g.tec Unicorn, Neurosity Crown, NeuroSky, Ōura and others). If your device is not listed, it can still be used as long as it exposes a compatible LSL stream.
Are the foundation model weights downloaded automatically?
No. The quickstart asks you to install the Hugging Face CLI and run hf download manually (hf download Zyphra/ZUNA is the example). The selectable models are listed in src-tauri/exg_catalog.json, which includes ZUNA, LUNA, REVE, OpenTSLM, SensorLM, SleepFM, SleepLM, OSF, SignalJEPA, LaBraM, EEGPT, TRIBE v2, NeuroRVQ, CBraMod and ST-EEGFormer entries.
Can I run it as a background service on a server?
The source documents only the Tauri desktop build (npm run tauri build) and desktop release artifacts; there is no Docker or headless deployment guidance. For programmatic access, connect to the app's WebSocket/HTTP API as described in docs/API.md.
Is NeuroSkill — Local-First EXG/BCI Neurofeedback Workbench open source?
Yes. The source code is on GitHub (NeuroSkill-com/skill) under the GPL-3.0 licence. GPL-3.0 is a copyleft licence: if you distribute a modified version, it must stay under the same terms.
Is NeuroSkill — Local-First EXG/BCI Neurofeedback Workbench safe to use?
FollowAgents' FARS review (Nov 08, 2026) scored NeuroSkill — Local-First EXG/BCI Neurofeedback Workbench 13/29 on trust and safety. None of the eight safety controls the review checks for was missing outright. The repository has changed since that review.
Is NeuroSkill — Local-First EXG/BCI Neurofeedback Workbench still maintained?
Yes. The latest commit on GitHub is from 2026-11-17, within the last 30 days. The latest release is v0.1.2 (2026-11-13).
Who makes NeuroSkill — Local-First EXG/BCI Neurofeedback Workbench?
NeuroSkill — Local-First EXG/BCI Neurofeedback Workbench is published on GitHub by the account NeuroSkill-com. FollowAgents has not verified which person or company is behind that account.
What is NeuroSkill — Local-First EXG/BCI Neurofeedback Workbench's official website?
The project's GitHub repository lists https://neuroskill.com as its website. That link is set by the repository owner; the source code itself is on GitHub (NeuroSkill-com/skill).
Was this page helpful?
View on GitHub ↗ Install ↓

Related concepts