Every project, tool and team building neurotech in the open — discovered automatically, refreshed every 3 hours, relevance-filtered, and free of noise.
The radar is generated from public GitHub topic search. A repository is kept if it carries a core BCI/neuro topic, or a context topic (signal processing, embedded, privacy…) plus an anchored neuro keyword — so generic projects don't leak in. Recency is measured from a fixed 6-hour snapshot, so unchanged projects keep a stable position. It's a discovery map — not an endorsement or a ranking of quality, safety, or clinical fitness. AxonOS is ranked by the same formula as everyone else.
Position is a neutral discovery signal: log₁₀(stars+1) × 10 + recency. Rising reflects real 7-day star velocity from accumulated history. No hand-tuned “hype”, no paid placement, no self-boosting — every project, including AxonOS, is ranked by the same formula.
Each card carries a Health score (0–100) computed only from real public GitHub signals the scan already fetches — no extra calls, no surveys, no self-assessment. Six measurable dimensions feed a weighted mean, and any dimension we can’t measure for a repo is left out rather than guessed:
We deliberately do not score conformance, security posture or test quality: none can be read reliably from search metadata, so scoring them would be fiction. A card marked search-only means the repo wasn’t enriched this cycle, so momentum and team were not measured. Health is a signal to help you triage the field — never a guarantee of quality, safety, or clinical fitness, and AxonOS is scored by the exact same rules as everyone else. Full formula: docs/METHODOLOGY.md.
The dataset refreshes automatically every 3 hours and is published to a single, always-current live stats issue. You can also publish a fresh scan to that issue on demand:
Maintainers: the button opens the scan workflow — click Run workflow to rescan GitHub, refresh the data, and rewrite the live stats issue in place.The question the field actually asks before adopting a dependency is not "how many stars" but "can I build on this?" — will it cite, will it take a vulnerability report, does anything verify it? The Foundation read-out on each card answers with seven checkable repository facts — files and workflows that either exist or don't. It is not a quality score and it never guesses:
Honest limits, stated plainly: we check the root-level SECURITY.md and the exact CITATION.cff filename (GitHub's own citation mechanism); the first four facts come from GitHub's community profile. When the evidence is indeterminate this cycle — a rate limit, a transient error — the card simply shows no Foundation read-out rather than a fabricated zero. And a missing file says a file is missing — it says nothing about code quality.
Open BCI's most expensive pain is integration guesswork: every stack has its own transport and format, and finding "everything that speaks LSL" means an afternoon of tab-hopping. The radar now detects interop tags from each project's topics, description and homepage against a committed vocabulary — data/interop-vocab.json — and makes them a one-tap filter ("Speaks") above the grid.
The vocabulary covers protocols (LSL, BrainFlow), data formats (BIDS, EDF), analysis toolkits (MNE, EEGLAB, OpenViBE, Timeflux), hardware families (OpenBCI, Emotiv, Neurosity, g.tec, Muse, ADS1299, FreeEEG) and runtimes (ROS, Unity, Arduino, ESP32, BLE).
Detection is deliberately conservative: word-boundary matching only, so unity never fires inside community and mne never fires inside mnemonic; collision-prone words require qualified forms (ros2, muse-lsl). It is heuristic — it reads what maintainers wrote, so a project that never mentions its protocol won't be tagged. Tags mean "this project references the ecosystem", not "certified compatible".
Everything on this page is backed by public, machine-readable endpoints you can build on — same origin, CORS-friendly, refreshed with every scan or deploy:
Provenance you can verify: the scan bot writes data through the GitHub API, so every data commit carries GitHub's Verified signature; the payload passes a fabrication-guarding validator and the published JSON Schema in CI; and a documentation-coverage gate fails the build if any shipped signal, tag or endpoint is missing from this very page. Nothing on a card is unexplained.
AxonOS is an open, real-time neural operating system for brain–computer interfaces — a no_std Rust kernel, a wire-level consent protocol, a deterministic signal pipeline, and the standard that ties them together. This is what's shipped, in progress, and planned.
Real-time no_std kernel: deterministic scheduler, SPSC queues, monotonic time, capability model.
The consent state machine, with six machine-checked proofs over grant, revocation and expiry.
The wire format: one encoding, decoded byte-identically in five languages.
Deterministic fixed-point conditioning. Same input, same bytes, any machine.
The specification the RFCs are tracked against.
Byte-exact vectors for RFC-0005, 0006, 0008 and 0009, re-derived from published constants rather than transcribed.
This radar: an auditable map of the open BCI field, rescored every three hours.
The scoring rule itself, published so any score on the radar can be recomputed from public data.
The contract with silicon: a timing budget that refuses, and re-closes at every operating point.
Nine RFCs. The two in draft say why they are draft.
Ask anything about AxonOS. Every answer carries its source, or says it does not have one.
The cards above are a snapshot. The full public roadmap — every milestone for the radar and the wider AxonOS project, with live Todo / In progress / Shipped status — is tracked in the open on the project board. Nothing sits behind a private tracker.
📋 Open the AxonOS roadmap board →The open BCI field is scattered. Pick the path that fits you — every link goes to something real.
You want datasets, decoders and tooling that actually exist.
You want to build — or find a codebase worth contributing to.
You want one honest map instead of a hundred browser tabs.
You have an open BCI/neuro project the map should know about.
Every planned feature for the radar and the wider AxonOS project — what's shipped, what's in progress, and what's next — lives on one public board. Nothing is hidden behind a private tracker: milestones, RFCs and the release pipeline are visible to anyone who wants to follow along or pitch in.
Open the project board →The radar is the community edge of a larger open project. AxonOS is an open, real-time neural OS for BCIs — explore the engineering at axonos.org and the code on GitHub.