wandb/server

W&B Server is the self hosted version of Weights & Biases

What it solves

It provides a self-hosted version of the Weights & Biases platform, allowing teams to keep their machine learning experiment tracking and data on their own infrastructure rather than using a cloud-hosted service.

How it works

The system is deployed as a Docker container. Users can start the server using the wandb server start command or via a manual Docker run command. Once running, other machines in the network can report metrics to this server by configuring the host URL and API key.

Who it’s for

Machine learning engineers and researchers who require self-hosted infrastructure for their experiments, specifically those who need to maintain control over their data and environment.

Highlights

  • Self-hosted deployment via Docker.
  • Support for production-grade features like external MySQL, cloud storage, and SSO (with a license).
  • Provisioning scripts available for AWS, Google Cloud, and Azure via Terraform.

Related

  • Project
    wandb/wandb

    A machine learning experiment tracking and visualization platform that helps developers monitor hyperparameters and metrics across the entire ML pipeline.

  • Project
    wandb/examples

    A curated set of ready‑to‑run scripts and Colab notebooks that show how to plug the Weights & Biases tracking library into popular ML frameworks (PyTorch, TensorFlow/Keras, Hugging Face, XGBoost, scikit‑learn, etc.). It demonstrates logging runs, configs, metrics, gradients, and artifacts, helping users add experiment tracking and reproducibility to their models with just a few lines of code.

  • Project
    HarnessRouter/harnessrouter

    A self-hosted platform for running AI agent harnesses locally via Docker, providing secure isolation and support for multiple model providers without cloud accounts.

  • Project
    gradio-app/trackio

    A lightweight, local-first experiment tracking library that provides a free alternative to Weights & Biases, featuring a Gradio-inspired dashboard and SQL-based querying for AI agents.

  • Project
    google-deepmind/xmanager

    A framework for packaging, running, and tracking machine learning experiments across local environments, Kubernetes, and Google Cloud Platform.