alexeymlartisan/mlartisan-selfhosted repository overview

⁠Quick reference

• Where to get help: [email protected]⁠
• Cloud version: https://mlartisan.com⁠

⁠MLArtisan - Self-Hosted

MLArtisan⁠ is a self-hosted solution for managing the full lifecycle of machine learning models.
It provides an API-driven platform that allows teams to quickly integrate, train, and run ML models without spending time on infrastructure setup or custom integration.

⁠Key Features

• Schema Management – Define datasets with clear feature/target structure.
• Model Training – Train using multiple algorithms (classification, regression, clustering, neural networks, etc.).
• Versioning – Store and manage different versions of models.
• Running Models – Send new data for predictions through a simple API.
• Self-Hosted – Full control of your data and compute environment with Docker.

⁠Use Cases

• Quickly add machine learning to existing projects without re-implementing training pipelines.
• Manage datasets and trained models in one place.
• Deploy models in secure, private environments.

⁠Deployment

Run MLArtisan in your own environment using Docker:

1. Pull the Docker Image:

docker pull alexeymlartisan/mlartisan-selfhosted:latest

1. Prepare your PostgreSQL database.
2. Run the Container (or configure env variables outside of docker run):

docker run -d -p 80:80 \
  -e ConnectionStrings__DefaultConnection="Host=your-host;Database=your-db;Port=5432;Username=your-user;Password=your-pass;Ssl Mode=Require;Trust Server Certificate=true;" \
  -e AccessTokenSecretKey="your-very-strong-secret-key" \
  -e SelfHostedService__IsLoginEnabled=true \
  -e SelfHostedService__Username="admin" \
  -e SelfHostedService__Password="admin-pass" \
  -e Instance__Host="http://your-url.com" \
  -e Instance__Type=2 \
  alexeymlartisan/mlartisan-selfhosted:latest

1. Open in Browser and visit http://your-url.com⁠ to start building models.

⁠Configuration Validation

The application requires several configuration values to be set before startup.
If any of them are missing, the service will throw an exception during initialization.

⁠Required environment variables

• Database (PostgreSQL)

  • ConnectionStrings__DefaultConnection must be provided.
    Example format:

    Host={{Host}};Database={{DBName}};Port=5432;Username={{Username}};Password={{Password}};Ssl Mode=Require;Trust Server Certificate=true;
    

    Copy

• Authentication

  • AccessTokenSecretKey must be set.

    • Used as the signing key for generating JWT tokens.
    • Important: choose a strong, random secret key (at least 32 characters).

  • At least one authentication method must be enabled:

    • Self-Hosted Login
      • SelfHostedService__IsLoginEnabled = true requires:
        • SelfHostedService__Username
        • SelfHostedService__Password
    • Google Authentication
      • Google__IsEnabled = true requires:
        • Google__ClientKey
        • Google__Secret
        • See Google Identity Platform Docs⁠ for setup instructions.

• Instance

  • Instance__Host → the public URL of this app instance (used by clients and other services).
  • Instance__Type → defines the role of the instance:
    • 0 = Trainer (only used for training models)
    • 1 = Runner (only used for serving predictions)
    • 2 = Both (training and running in the same instance)

  This allows running multiple coordinated instances:

  • One instance dedicated to training
  • Another dedicated to running models

⁠🔒 Behavior

If any required variable is missing or misconfigured, the application will not start and will throw a descriptive exception.

⁠Terms and Privacy

By using this service, you agree to the Terms⁠ of Use and Privacy Policy⁠.