Compiler, not dashboard

Battery model passports, compiled from real test data.

BATT-MX turns battery cell definitions, cycling CSVs, test protocols, degradation models, validation rules, uncertainty, and provenance into portable .battx battery model passport packages.

Request technical preview View sample passport

Input

.battlang, sealed .batt, cycling CSVs

Output

.battx package, report, normalized data

Scope

Evidence infrastructure for passport readiness

Compiler pipeline

.batt source

Sealed compiler-readable input derived from editable .battlang.

CSV cycling data

Normalized from Arbin, Neware, Maccor, BioLogic, and lab exports.

BATT-MX compiler

Validates protocol, model fit, uncertainty, and provenance.

.battx passport

Portable model package with reports, manifests, exports, and hashes.

Preserves protocol, provenance, validation, and uncertainty from source files through compiled package output.

Regulatory timing

Building the technical evidence layer for battery passport readiness.

The EU Battery Passport requirement is approaching for EV, LMT, and industrial batteries over 2 kWh placed on the EU market. Companies will need structured, accurate, traceable battery records.

Important boundary

BATT-MX does not claim full EU compliance today.

Current focus

Compiler artifacts that preserve data lineage, model assumptions, validation results, and package structure.

Readiness layer

Designed to sit underneath digital battery passport, battery analytics, and compliance workflows.

Problem

Battery evidence is trapped in CSVs, notebooks, PDFs, and undocumented assumptions.

Cycler exports are inconsistent across Arbin, Neware, Maccor, BioLogic, and lab formats.

Protocol context is often separated from data.

Degradation models are hard to audit.

Supplier claims are difficult to validate.

Compliance teams need traceability, not screenshots.

Engineers need model provenance, not another dashboard.

Product

A compiler for battery truth.

BATT-MX separates editable source, sealed compiler input, validation logic, normalized data, and compiled passport output.

BattLang

Proprietary battery-domain source format for cell definitions, protocols, datasets, model fit instructions, and validation rules.

BATT-MX

Compiler and validator that seals inputs, normalizes data, checks assumptions, and emits portable evidence packages.

.battx

Compiled battery model passport package for model evidence, normalized data, validation reports, provenance, and exports.

sample .battlang source

model[degradation_model][
    cell[LG_M50T]
    data[cycle_test_01]
    protocol[aging_test]
    mechanisms[sei_growth, lithium_plating, loss_of_active_material]
    fit[
        target[capacity_fade]
        cycles[1..500]
        method[baseline_curve_fit]
    ]
    validate[
        holdout_cycles[501..700]
        max_error[3.5 %]
    ]
]

.battlang

Editable source for battery-domain definitions and model instructions.

.batt

Sealed compiler-readable input produced from .battlang.

.battx

Compiled passport package containing evidence, reports, data, provenance, and exports.

Technical output

What comes out of a .battx package.

A .battx package is designed to travel with the model evidence, not just point at a dashboard.

.battx package tree

LG_M50T_model.battx/
  metadata.json
  cell.json
  dataset_manifest.json
  protocol.json
  model.json
  degradation_summary.json
  uncertainty.json
  validation_report.md
  normalized_data.csv
  provenance.json
  exports/
    pybamm_model.py
    fmu_stub.txt

metadata.json

Passport identity, schema version, compiler version, timestamps, and package notes.

cell.json

Cell chemistry, nominal values, manufacturer fields, and declared test context.

dataset_manifest.json

Dataset list, source files, row counts, hashes, units, and normalization status.

protocol.json

Cycling protocol, temperature, charge/discharge steps, rest periods, and holdout ranges.

model.json

Selected model family, fitted parameters, target variable, and validation scope.

degradation_summary.json

Capacity fade, dominant mechanisms, selected curve form, and bounded conclusions.

uncertainty.json

Confidence notes, validation range, RMSE, missing data, and known assumptions.

validation_report.md

Human-readable validation report with pass/fail checks and evidence gaps.

normalized_data.csv

Normalized cycling data with canonical columns and unit handling.

provenance.json

Source file hashes, transformation log, compiler inputs, and reproducibility metadata.

exports/pybamm_model.py

Planned PyBaMM export surface for downstream modeling workflows.

exports/fmu_stub.txt

Placeholder for future FMU export through a real FMI toolchain.

Sample passport

Validation report preview.

Synthetic sample shown for product communication. Public benchmarks and real importer examples are on the roadmap.

sample compile output

Model: degradation_model
Chemistry: NMC811
Selected model: exponential_fade
Validation RMSE: 0.0019
Confidence: high
Included files: metadata.json, cell.json, dataset_manifest.json, protocol.json, model.json, degradation_summary.json, uncertainty.json, validation_report.md, normalized_data.csv, provenance.json

Validation RMSE

0.0019

Within declared holdout range.

Confidence

High

Based on synthetic sample coverage.

Uncertainty

Included

Assumptions and missing data travel with the package.

Use cases

From messy cycling data to validated model evidence.

Battery labs

Turn test campaigns into client-ready model evidence packages.

Cell manufacturers

Package validated performance claims with source data.

Pack integrators

Compare supplier cells with normalized validation artifacts.

Research groups

Make degradation model datasets reproducible.

Compliance teams

Prepare technical evidence for digital battery passport systems.

Trust and limitations

Evidence infrastructure, not safety certification.

BATT-MX is not a safety certification tool. It packages auditable engineering evidence.

Validation range is explicit.

Assumptions are included.

Missing data is reported.

Source files are hashed.

Uncertainty is included.

Future packages can be signed.

No one-click compliance or safety guarantee is claimed.

Investor view

Infrastructure for the battery passport era.

BATT-MX occupies the technical evidence/compiler layer underneath battery analytics, digital product passport, and compliance workflows.

Signal

Regulatory tailwind from EU battery passport requirements.

Signal

Battery data fragmentation across labs and suppliers.

Signal

Proprietary source format and compiler workflow.

Signal

Enterprise licensing potential.

Signal

Expansion path into importers, compliance exports, signatures, PyBaMM/FMU interoperability.

Signal

Can integrate underneath existing battery passport and battery analytics platforms.

Roadmap

Near-term engineering milestones.

Importer coverage

Arbin, Neware, Maccor, and BioLogic importers.

Passport schemas

EU 2027 passport evidence schema mapping.

Package integrity

Signed .battx packages.

Readiness checks

Compliance readiness checker.

Model exports

Real PyBaMM export.

Interoperability

FMU export through a real FMI toolchain.

Benchmarks

More public benchmark datasets.

Docs

Short technical path.

The first docs cover BattLang syntax, the compiler pipeline, package format, limitations, and CLI usage.

CLI quickstart

battmx seal examples/lg_m50t.battlang examples/lg_m50t.batt
battmx validate examples/lg_m50t.batt
battmx compile examples/lg_m50t.batt
battmx inspect build/LG_M50T_model.battx

Open technical docs

Technical preview

Testing with battery labs, researchers, and pack teams.

BATT-MX is seeking technical preview conversations with teams that have cycling data, validation requirements, and a need for portable model evidence.