A reproducible shelf-life decision, with the model record attached.

OpenPharmaStability analyzes CSV or XLSX stability data locally in Python using ICH Q1E-inspired fixed-effect batch regression and one-sided confidence-bound crossing.

Inspect the 17-month example Read the method

$pip install openpharmastability

Supported shelf life17 months

Governing batchB2

Modelcommon slope + batch intercepts

Statistical crossing17.954842 months

Slope interaction p-value0.9056

Batch intercept p-value1.658e-16

HTML reportreport.html PDF reportreport.pdf JSON decision recordrecord.json Confidence plotplot.png

The decision is visible in the evidence.

Golden assay example: 42 rows, 3 batches, 7 time points from 0 to 24 months. The lower one-sided 95% confidence bound crosses the 95.0% specification at 17.95 months.

Open the complete report

Assay confidence-bound plot for three batches showing a 17.95-month crossing of the lower specification limit — Real output from the Python engine—not a decorative dashboard. The supported shelf life is rounded down to 17 months; B2 is the governing batch. Observed coverage extends to 24 months, so the result does not depend on extrapolation.

Method record

Engine	Python statistical engine; local execution
Model	Fixed-effect batch regression; common slope with batch-specific intercepts
Poolability	Slope interaction p=0.9056; batch intercept p=1.658e-16
Decision rule	One-sided 95% mean-response confidence bound crossing the lower specification limit
Observed data	42 rows; 3 batches; time points 0, 3, 6, 9, 12, 18, and 24 months
Outputs	HTML report, JSON decision record, confidence plot, and optional PDF
Multi-attribute sample	impurity_a limits the product decision at 7 months

Supported scope

CSV and XLSX stability data
Fixed-effect multi-batch regression
One-sided confidence-bound crossing
Multi-attribute limiting decisions
Inspectable report artifacts

Explicit boundaries

Regulatory approval or submission
Validated GxP operation
21 CFR Part 11 controls
Automatic scientific judgment

Local execution

Your data stays on your machine. The public site is documentation and sample evidence; analyses run through the installed Python package and local workspace.

Decision support only

OpenPharmaStability provides statistical evidence for qualified review. It does not replace scientific judgment, quality review, or regulatory decision-making.

From stability data to an inspectable record.

The browser presents the evidence. The Python package owns the calculations.

Prepare data

Provide batch, condition, time, attribute, value, and specification columns in CSV or XLSX.

Run locally

The engine validates the data, tests batch poolability, fits the model, and computes the bound.

Review evidence

Inspect the plot, assumptions, warnings, model choice, and machine-readable record.

Use professional judgment

Bring the evidence into a qualified stability review without overstating the software's regulatory role.

Install and launch the local workspace

$ pip install openpharmastability
$ openpharmastability-ui --host 127.0.0.1 --port 8765

Run the golden example from the CLI

$ openpharmastability analyze examples/assay_3batch.csv \
    --condition "25C/60RH" \
    --attribute assay \
    --output report.html