BUILD YOUR OWN

Your code.
Your models.
Our plumbing.

You want to own your battery analytics. We respect that. But writing and maintaining parsers for every cycler format in your lab isn't where your value is — it's a tax on getting to the work that matters.

The Micantis Data Layer handles ingestion from every major cycler format, normalizes it, keeps it in the cloud, and gives you a Python API with Parquet downloads. Just clean data, ready for whatever you want to build.

This is the API-only tier. There is no user interface — you work entirely through the Python API and Parquet downloads. Your code, your notebooks, your dashboards. When your team is ready for a full platform with built-in analysis, dashboards, and reporting, upgrading to Micantis WorkBook is straightforward — your data layer stays the same.
$1,500 /month
Book a 15-min walkthrough → 
# Connect to your data
from micantis import MicantisAPI
api = MicantisAPI()

# Every cycler. Already normalized.
data = api.get_data_table(
    barcode="LOT-2024-0847"
)
df = api.download_parquet_file(data[0]["id"])

# Cycle summaries are already computed
df.columns
# → ['cycle', 'capacity_ah', 'energy_wh',
#    'dcir_mohm', 'voltage_v', 'current_a',
#    'temperature_c', 'timestamp', ...]

# Your models. Your dashboards. Your IP.
predictions = your_model(df)
⟵⟶

Every Format

Every major cycler format ingested automatically. No custom parsers. No maintenance when firmware updates.

☁

Cloud Native

Your data lives in the cloud, normalized and structured. Access it from anywhere. No local file management.

{ }

Python API

Clean, documented API. Pull any batch, any lot, any date range. Integrate with your existing tools and workflows.

▤

Parquet Out

Download as Parquet files. Fast, columnar, and ready for pandas, Spark, or whatever your stack needs.

Automatic Data Ingestion

Upload from any cycler. We handle the parsing, normalization, and storage.

Supported Cyclers

Every major format, ingested automatically. No custom parsers. No maintenance when firmware updates break your pipeline.

Arbin Maccor Neware Bitrode BioLogic Novonix PEC Gamry Basytec Landt 50+ more

What Happens On Upload

  • Auto-detect cycler format and firmware version
  • Normalize column names, units, and timestamps
  • Compute cycle summaries (capacity, energy, efficiency, DCIR)
  • Extract metadata and link to cell records
  • Store raw + processed data in cloud

Parquet Downloads

Columnar, fast, and ready for your stack. Cycle summaries are pre-computed and embedded.

# Download a full dataset as Parquet
df = api.download_parquet_file(data_id)

# Or filter by cycle range
df = api.download_parquet_file(
    data_id,
    cycles="1-50"       # first 50 cycles
)

# Last 10 cycles
df = api.download_parquet_file(
    data_id,
    cycles="-10"       # from the back
)

# Embedded metadata comes with the file
meta = api.unpack_parquet(df)
# → cell info, timestamps, cycle counts

What's In The Parquet

  • Raw time-series — voltage, current, temperature, timestamps
  • Cycle summaries — capacity, energy, Coulombic efficiency, DCIR (pre-computed)
  • Embedded metadata — cell ID, test name, cycler info, date ranges
  • Flexible ranges — download full datasets or specific cycle windows

Works With Everything

Parquet is the standard for analytical workloads. Load into pandas, Polars, Spark, DuckDB, Databricks, or any tool that reads columnar data. No proprietary formats.

Search, Filter & Metadata

Find what you need. Tag cells with your own properties. Build your data pipeline around the API.

Find Data

Search and filter your data table by barcode, date range, station, channel, cell test name, or data type. Get back structured results you can iterate over.

Per-Cell Metadata

Read and write custom metadata on any cell. Tag with chemistry, supplier, lot number, grade — whatever your workflow needs. Query it back in wide format for batch analysis.

Data Operations

  • stitch_data() — combine interrupted tests
  • clean_data() — auto-fixup, filter, parametric cleaning
  • get_changelog() — track every data modification
  • get_duplicate_files() — find similar uploads
# Search your data
results = api.get_data_table(
    search="NMC811",
    start_date="2024-01-01",
    end_date="2024-06-30"
)

# Read per-cell metadata
meta = api.get_cell_metadata(
    cell_ids=["CELL-001", "CELL-002"],
    metadata=["chemistry", "supplier", "lot"]
)

# Write your own properties back
api.write_cell_metadata([{
    "cell": "CELL-001",
    "field": "grade",
    "value": "A"
}])

# Track changes since last sync
changes = api.get_changelog(
    since="2024-06-01T00:00:00Z"
)

What You Keep

This is a foundation, not a platform. Everything above the data layer is yours.

Your Analysis

Write your own models, your own statistical tests, your own degradation fits. The data layer doesn't care what you do with the data.

Your Dashboards

Build in Grafana, Streamlit, Jupyter, or whatever your team already uses. Pull from the API. We're not replacing your tools.

Your IP

Your models, your code, your insights — they stay yours. We provide the plumbing. You own everything that runs on top of it.

Install from PyPI
pip install micantis[parquet]
View on PyPI →

Ready for More?

The Data Layer is the foundation. When your team needs more, the upgrade is seamless.

Same Data Layer

Upgrading to Micantis WorkBook doesn't change your data infrastructure. Everything you've built on top of the API keeps working.

Add a Full Platform

WorkBook adds dashboards, quality control workflows, automated reporting, real-time monitoring, and built-in analysis tools — all connected to the same data.

Add MOOSE

Ask questions about your test data in plain language. Get charts, key findings, and scheduled reports delivered to your inbox. Learn more →

Skip the first three months. Start where the data is already clean.
15-minute walkthrough. See the API. Ask anything.
Book a Walkthrough →