Results

Results you can inspect

Explore the published benchmarks and download an NDC compile report with its validation, model structure and checksums.

Public download: eight report files. Input trajectories and executable model files are not included.

Benchmarks

What the engine did on public data

+8.2 dB
on multi-tone vibration at 20 dB SNR — where the best classical filter goes to −9.3 dB.
6/6
free-run wins for compiled models against classical methods on public cases.
32/32
real-waveform cases carried to the end, including the ones where the engine declined and printed it.
0
universal performance guarantees from these examples.

These are engine-level results on signal benchmarks. The battery pack is a separate table with its own dataset and its own baselines, and it will be published the same way: dataset, protocol, both outcomes.

Artifact

Download the published NDC reports

This ZIP contains the eight published report and provenance files from the R9 NDC parity example, with a checksum list for this public subset. The input trajectories and runnable model files are not included. The original manifests describe the larger delivery.

Check the public-subset hashes after extraction. This establishes byte integrity of these files, not reproduction of model predictions or independent model accuracy.

Download NDC report ZIP

# Extract ndc-public-reports.zip, then check its eight original files:
# macOS:
shasum -a 256 -c PUBLIC_SUBSET_CHECKSUMS.sha256
# Linux:
sha256sum -c PUBLIC_SUBSET_CHECKSUMS.sha256
# These commands verify files, not model quality or predictions.

How we word our claims

Three rules the whole site is written by

01

Every number is a run

Performance figures belong to a stated test, version and metric. Distinguish observed results, supported product limits and illustrative examples; none should be presented as a guarantee for every input.

02

A refusal is a result too

Identity/pass-through is an explicit result status. A numerical gain of 0 dB alone does not prove identical input and output or establish why processing was withheld. Read the recorded status and protocol.

03

Every conclusion has a boundary

Each model records the observed operating range and its policy for unsupported inputs. Follow the policy in the delivered artifact. A value outside validated conditions is not a validated prediction.

Comparison

Where we are strong, and where to pick something else

Your own scriptAn enterprise platformnlsys
Cost of entryYour own timeCheck the provider’s purchase routeAccount access; introductory allowance shown in the workspace
ScaleDepends on the implemented pipelineDepends on the product and deploymentFrom a single file to a campaign
FormatsWhatever you writeCheck supported exports, units and mapping requirementsCSV and TSV through a mapping step
Record of changesWhatever you logCheck the granularity and exportability of change recordsProcessing and integrity records. Per-value details are specific to the selected Battery output.
Behaviour when unsureHowever you code itCheck documented uncertainty and refusal behaviorInspect readiness, support and refusal statuses in the delivered output
What you keep if you leaveYour codeCheck export, licence and retention termsDownloaded files, subject to their licence; see the deliverable and verifier scope

What we are good at

  • A result that survives review: you can show not only the outcome but what was done to the data.
  • Battery feature claims require the corresponding battery-specific test and reported acceptance criteria, not generic denoising scores.
  • Start with a free local file screen, then use the relevant account workflow.
  • Keep downloaded files under the delivered licence. Verify which runtime and tools are included; hosted retention is a separate condition.

What we do not do

  • We do not read binary .ndax, .mpr and .res — we work with CSV and TSV. How to export.
  • We do not replace a LIMS, and we do not do supplier scorecards, sampling plans or lab scheduling.
  • We do not issue the regulatory EU Battery Passport under Regulation 2023/1542: our dossier is an engineering record of an experiment.
  • We do not sell autonomous actuation: Control computes setpoints, your system executes them.

Start

Check it on your own data

The local browser check is free and needs no account. Confirm columns and units, review its findings, and select the product appropriate to your task.