Self Hosted

Self Hosted installations for projects with sensitive data

W&B Local is the self hosted version of Weights & Biases. It makes collaborative experiment tracking possible for enterprise machine learning teams, giving you a way to keep all training data and metadata within your organization's network.

Request a demo to try out W&B Local →

We also offer W&B Enterprise Cloud, which runs a completely scalable infrastructure within your company's AWS or GCP account. This system can scale to any level of usage.


  • Unlimited runs, experiments, and reports

  • Keep your data safe on your own company's network

  • Integrate with your company's authentication system

  • Premier support by the W&B engineering team

The self hosted server is a single Docker image that is simple to deploy. Your W&B data is saved on a persistent volume or an external database so data can be preserved across container versions.

Server Requirements

The W&B self hosted server requires an instance with at least 4 cores and 8GB memory.

Self Hosted Resources