Spike sorting, done right — and reproducibly.
Automated spike sorting free; probe-specific models, manual curation, and cluster-quality metrics at Pro. On your own machine, in open formats, with quality and provenance on every unit.
Sorting is where results are made — or lost.
Spike sorting decides what counts as a unit, so a fragile or inconsistent sort quietly compromises everything downstream. Most labs stitch it together from scripts and one-off tools — no quality metrics, no way to re-derive the result. Radiens does the standard sort the same way every time — automated by default, curatable when you need it — and stamps quality and provenance onto every unit, so the result re-derives itself and hands cleanly to Python, open standards, or an agent.
Automated by default, professional when you need it.
Sort live during acquisition in Allego, post-hoc in Videre, or from your own code — the same engine underneath. Standard-model sorting is free; Pro adds the probe-specific models and curation that publication-grade work needs.
Automated, standard-model sorting — free
One-click automated sorting in Videre and live in Allego; spike-train analysis and a per-unit view included.
Probe-specific models + manual curation
Probe-specific sorting models, plus manual curation for the units that need a human.
Cluster-quality metrics + unit tracking
Cluster-quality metrics on every unit, plus cross-session tracking with drift correction.
What you get, by tier
The full sorting stack scales with your work — free to start, professional at Pro, cluster-scale on the roadmap.
- ✓Standard (free) — automated spike sorting, live and post-hoc, with spike-train analysis.
- ✓Pro — probe-specific models, manual curation, cluster-quality metrics, and cross-session unit tracking.
- ✓On the roadmap — GPU / cluster-scale sorting and model-based spike analysis (Radiens Compute).
Why it holds up
Yours, local-first
Sort on your own machine against your own storage — nothing uploaded.
Open in, open out
Read most vendor formats; export sorted units to NWB and CSV.
Provenance on every unit
Each result carries a unique ID and provenance, so it re-derives itself.
Native probe models
Probe-specific sorting for NeuroNexus probes and both CMOS active-probe families, SiNAPS and Neuropixels.
Frequently asked questions
Does Radiens include spike sorting for free?
Yes — automated standard-model spike sorting is in the free Standard tier, live in Allego and post-hoc in Videre, with spike-train analysis and a per-unit view. Pro adds probe-specific models and manual curation.
What automated spike sorting does Radiens use?
A standard sorting model, one-click, run the same way every time so results stay reproducible. Probe-specific sorting models are a Pro upgrade.
Can I manually curate sorted units in Radiens?
Yes, at Pro: split, merge, and refine units by hand on top of the automated sort, with the changes carried in the provenance record.
What cluster-quality metrics does Radiens provide?
Isolation distance, L-ratio, and ISI on every cluster (Pro), plus signal-metric statistics to triage and rank units before you rely on them.
Does Radiens do probe-specific spike sorting for Neuropixels and SiNAPS?
Yes (Pro): probe-specific sorting models with native support for NeuroNexus probes and both CMOS active-probe families, SiNAPS and Neuropixels.
Can I run spike sorting from Python or the command line?
Yes — the same sorting is scriptable through RadiensPy (a free Python client) and the Radiens CLI, so it composes into pipelines and agent workflows.
Does Radiens do large-scale or GPU spike sorting?
On-demand cluster-scale and GPU spike sorting are in development for Radiens Compute, the programmable backend; today, sorting runs on your own workstation.