Installation - Synthefy Docs

Synthefy ships two packages. Pick by what you want to do:

I want to…	Install	Import
Run Nori locally (open-source, GPU or CPU)	`pip install synthefy-nori`	`from synthefy_nori import NoriRegressor`
Call hosted Nori (no GPU)	`pip install synthefy`	`from synthefy import SynthefyNoriClient`
Time-series forecasting	`pip install synthefy`	`from synthefy import SynthefyAPIClient`

synthefy-nori is the open-source tabular model. It downloads pretrained weights from Hugging Face and runs locally with a scikit-learn–style API. Start here for tabular regression.
synthefy is the client for the hosted platform — hosted Nori and time-series forecasting.

Both give you Nori. Use synthefy-nori to run the model yourself; use synthefy to call it (or the forecasting API) over the network.

Optional extras

Install only what a task needs — these are add-ons to synthefy-nori:

Command	Adds	For
`pip install "synthefy-nori[interpretability]"`	shapiq	SHAP / feature attribution
`pip install "synthefy-nori[eval]"`	matplotlib, openml	reproducing the benchmarks
`pip install "synthefy-nori[train]"`	wandb, xgboost	training from scratch

git clone https://github.com/Synthefy/synthefy-nori
cd synthefy-nori
uv sync --extra dev

Fit and predict in five lines.