OpenModelStudio

The Open-Source AI Laboratory

Build Models

Build, train, and deploy ML models on CPU or GPU with Kubernetes-native infrastructure. Multiple workspaces, one platform — runs anywhere.

Get Started
Join Discord
Terminal — zsh

Everything You Need to Build AI

From data exploration to production deployment, OpenModelStudio covers the entire ML lifecycle.

For Data Scientists

  • Project Management with stage-based workflow
  • In-browser Model Editor (Monaco, Python + Rust)
  • Model Registry — search, install, and train pre-built models
  • Real-Time Training with live loss curves via SSE
  • 9 Visualization Backends with drag-and-drop dashboards
  • Experiment Tracking with parallel coordinates
  • Global Search across models, datasets, and experiments
  • Multi-IDE Workspaces — JupyterLab, VSCode, and more
  • LLM Assistant for natural language control
  • AutoML with hyperparameter search

For ML Engineers

  • Kubernetes-Native — every model in its own pod
  • CPU or GPU — runs on any hardware, anywhere
  • Python SDK & CLI — pip install openmodelstudio
  • Rust API — high-performance Axum + SQLx backend
  • GraphQL — auto-generated from PostgreSQL
  • Real-Time Notifications — training events, job completion
  • Streaming Data — never load full datasets to disk
  • One-Command Deploy — make k8s-deploy

For Researchers

  • Reproducible Experiments — full session snapshots
  • PyTorch + Rust-Torch — write models in either
  • Consumer Hardware — train on laptops, not GPU clusters
  • Pre-Loaded Tutorial Notebooks in every workspace
  • Artifact Versioning tied to git commits
  • Dataset Streaming — work with massive datasets locally
  • Collaborative Workspaces — share and iterate together

Kubernetes-Native Architecture

Every component runs in its own pod. Models train in ephemeral jobs. Data streams — never stored on disk.

Kind Cluster
Frontend
Next.js + shadcn/ui
Rust API
High-performance Axum
PostGraphile
Auto-generated GraphQL
Workspaces
Jupyter, VSCode, and more
|
PostgreSQL
System of record
Model Runners
Ephemeral CPU/GPU pods

Training Job Lifecycle

Train
User triggers
K8s Job
API creates
Pod
Model trains
SSE
Metrics stream
Charts
Live updates
Artifacts
Results saved

See It In Action

A modern, beautiful interface for every step of your ML workflow.

OpenModelStudio — Dashboard
OpenModelStudio Dashboard and JupyterLab

Build or install models, track experiments, and deploy for inference.

OpenModelStudio — Workspaces
OpenModelStudio Workspaces and Model Metrics

Robust Visualization Framework for training and inference.

OpenModelStudio — Visualizations
OpenModelStudio Visualizations and Dashboards

Model Registry

Browse, install, and train pre-built models from the Open Model Registry. Self-contained models that work out of the box — install and start training in seconds.

iris-svm

sklearn

SVM classifier for the Iris dataset. Multi-class classification with RBF kernel.

classificationtabularmulti-class

mnist-cnn

pytorch

CNN for MNIST digit classification. Conv2d layers with deep learning fundamentals.

computer-visionimage-classificationcnn

sentiment-lstm

pytorch

Bidirectional LSTM for text sentiment analysis with attention-weighted pooling.

nlptext-classificationsentiment
$pip install openmodelstudio && openmodelstudio install iris-svm
Browse All Models

Built With the Best

Modern, battle-tested technologies at every layer of the stack.

Next.js
React
Rust
Axum
PostgreSQL
Kubernetes
PyTorch
Docker
GraphQL
TypeScript
Tailwind CSS
shadcn/ui

Up and Running in Minutes

Three commands. That's it. Docker, Kind, and make handle the rest.

Terminal — zsh

Join the Community

OpenModelStudio is open source under GPL-3.0. Star us on GitHub, join our Discord, or contribute directly.

0%
Open Source
0
Docker Images
0+
API Endpoints
Star on GitHub
Join Discord