FlowMine Studio

Process mining with a real UI — installed with pip, running in one command.

pip install flowmine-studio
flowmine-studio

That's it. Your browser opens http://localhost:8080 — sign in with admin / admin. A demo event log (a loan-application process, 5,000 cases) is already analyzed: open the demo project and you'll see the process map, metrics, charts, and automatic insights with estimated financial effects.

What it does

• Upload an event log — CSV, Excel, Parquet, or XES — map the case / activity / timestamp columns in a guided wizard, and get a full analysis in minutes.
• Connect to a database — PostgreSQL / Greenplum, Oracle, or any SQLAlchemy URL — point at a table or write a query, and mine directly from the source.
• Process map — DFG, Heuristic, Alpha and Alpha+ miners, with Disco-style activity/path sliders, frequency and duration heat, and PNG export.
• Metrics — activity, transition, trace, case and resource tables, sortable and exportable to CSV.
• AutoInsights — bottlenecks, loops, ping-pongs, growing durations and irregularities found automatically, ranked by anomaly level, with financial effect estimates.
• Charts — case duration distribution, activity frequency, Pareto, Sankey, Gantt, and time-intensity heatmap.
• Users and roles — administrator / analyst / viewer, for a shared instance.

The interface is bilingual (Russian / English). The engine is the flowmine library.

Where your data lives

Everything is local: an SQLite database and analysis artifacts in your user data directory (shown at startup). Nothing leaves your machine — no telemetry, no external services.

Options

flowmine-studio --port 8888        # another port
flowmine-studio --host 0.0.0.0     # listen on the network (change the admin password!)
flowmine-studio --no-browser       # don't open the browser
flowmine-studio --no-seed          # skip the demo data

Environment variables (STUDIO_*) override any setting — see flowmine_studio/api/config.py.

Development

pip install -e .[dev]
make test          # pytest (SQLite; Postgres connector tests auto-skip without a server)
make lint
make dev-api       # backend on :8080
make dev-ui        # UI dev server on :5173 (proxies /api)
make ui-mock       # UI with built-in demo mocks, no backend needed

The UI is a React app in ui/; npm run build outputs into the Python package (flowmine_studio/api/static/) so the wheel ships it.

License

AGPL-3.0-or-later. © FlowMine Contributors.