The privacy-first MLOps workspace. Ingest data, engineer features visually, and train models without your data ever leaving your infrastructure.
Built for teams who want to focus on models, not glue code.
Drag, drop, and connect nodes to clean data and engineer features. No more spaghetti code in notebooks.
Run heavy training jobs in the background. Powered by Polars for high-speed processing and Celery for robust task management.
Go from experiment to production instantly. Serve predictions via a built-in REST API.
Import CSV, Parquet, or SQL datasets directly. Powered by Polars, Skyulf auto-detects column types and caches everything for fast, repeatable experiments.
Automated profiling, outlier detection, and causal discovery. Spot data quality issues instantly with our high-performance Polars-based engine.
The Feature Canvas lets you visualize your data transformations. Handle missing values, encode categories, scale features, and more—all with drag-and-drop nodes.
Don't guess which model is better. Track every hyperparameter, metric, and artifact. Visualize performance with interactive charts.
Once you're happy with a model, deploy it to a local inference endpoint. Verify predictions immediately with the built-in JSON editor before moving to production.
Spin up a local API endpoint to test your model against real data. Debug inputs and outputs in real-time without deploying to a remote environment.
Continuous monitoring of your deployed models. Detect data drift, concept drift, and prediction anomalies in real-time to ensure model reliability.
Teams working with sensitive data who can't upload to cloud services.
Train models on patient data or research datasets that can't leave your infrastructure due to GDPR or institutional policies.
Municipalities and public institutions working with citizen data that must stay on-premises for privacy and compliance.
Small teams who need ML tooling but want to avoid expensive SaaS subscriptions and keep data under their own control.
Everything you need to know about the platform and how it works.
Skyulf is a self-hosted MLOps platform. It lets you build, train, and monitor machine learning pipelines entirely on your infrastructure. Unlike cloud-based solutions, your data never leaves your servers—perfect for privacy-sensitive or regulated environments.
I named it Skyulf after two ideas. Sky is the open space above Earth, where the sun, moon, stars, and clouds live. Ulf means “wolf,” with Nordic roots, and the wolf is also a strong symbol in Turkic tradition. Together it fits the project: independent and community-driven.
Local-first means complete data sovereignty. No vendor lock-in, no surprise cloud bills, no data privacy concerns. You control where your models train, where your data lives, and who can access it. It's MLOps on your terms.
Cloud platforms require uploading your data to third-party servers and charge based on usage. Skyulf runs 100% on-premise or in your private cloud. No data leaves your network, no usage-based pricing, and no vendor dependencies. You own the infrastructure and the workflow.
Skyulf is currently in Active Alpha. It is perfect for internal tools, research, and experimentation. We are working towards a stable v1.0 for critical production workloads.
Yes. Skyulf is open source with a split license model: Apache 2.0 for the backend and GNU AGPLv3 for the frontend. This ensures the core remains easy to integrate, while UI improvements are shared back to the community.
Contributions are welcome! Check out our GitHub. Whether it's code, documentation, bug reports, or feature ideas—every contribution helps shape Skyulf.
To democratize AI development. We are building an "App Hub" where anyone can drag-and-drop to create powerful AI tools—from traditional ML to GenAI agents—without needing a PhD or a cloud budget.
For the backend (Apache 2.0), commercial use is free. For the frontend (AGPLv3), if you build a proprietary SaaS that modifies the UI, you may need a commercial license exception.
From ingestion to deployment, Skyulf keeps your experiments reproducible and private. Here’s a snapshot of how a local-first team moves from raw data to a deployed model.
Point to CSV, Parquet, or SQL sources. Schema gets detected automatically.
Drag nodes around the canvas to wire up transforms. Preview stats at each step.
Launch training runs in the background. Deploy the best model to a local API.
Powered by modern open source
Join the local-first MLOps revolution. Open source and free forever for local use.
Launch Skyulf