SwanLab is an Apache-2.0, self-hostable experiment-tracking and visualization tool for ML training — an open-source alternative to Weights & Biases.
Yes, SwanLab is open source under the Apache-2.0 license.
SwanLab has a free tier plus paid plans. The open-source version can be self-hosted with no license cost; you pay only for your own infrastructure.
Yes, SwanLab can be self-hosted (the source is available under the Apache-2.0 license).
SwanLab is an AI training-tracking and visualization tool. It logs metrics, hyperparameters, and hardware utilization during model training, then lets you compare runs and collaborate across a team — the familiar experiment-tracking workflow, positioned as an open-source alternative to Weights & Biases. It is a Python library and is available both as a self-hosted deployment and as a hosted cloud service (hence the freemium model).
It fits ML teams that want experiment tracking under their own control, or that prefer an Apache-2.0-licensed option to a proprietary SaaS tracker.