Model integration & comparison

One API for running, tracking and retrieving model results.

MIC3 is a central platform for starting model runs, tracking their status, reusing cached results and retrieving artifacts — across independently maintained modelling projects.

Start runs from one API

Every modelling project stays an independent container. One request shape starts any of them.

Track status reliably

Runs are recorded before they are scheduled, so pending, running, completed and failed are always answerable.

Reuse cached results

A completed run is reused when the model version, adapter, inputs, config and scenario parameters all match.

Index artifacts, not schemas

Spreadsheets, GeoJSON, NetCDF, SQLite, reports or logs — files are indexed and downloadable without a shared output schema.

For modelling teams

Adding a model to MIC3 does not mean rewriting it as a service. A team provides a container image, one non-interactive command that runs the model to completion, a machine-readable config the platform can fill in, a deterministic output directory, and a contract describing which artifacts that run produces.