timestep 13.0.0a7

Timestep AI CLI - free, local-first, open-source AI

Timestep AI

Timestep AI CLI - free, local-first, open-source AI

Overview

This project provides a modular infrastructure for data engineering, machine learning, and deployment pipelines.

Project Structure

project_root/
│
├── infra/                  # Infrastructure management
│   ├── cloud_management/   # Cloud instance operations
│   │   └── cloud_instance_controller.py
│   │       - Manages cloud instances using Apache Libcloud
│   │
│   ├── cluster_management/ # Kubernetes cluster management
│   │   └── k3s_cluster_controller.py
│   │       - Manages K3s Kubernetes clusters
│   │
│   └── workload_management/ # Workload orchestration
│       └── sky_workload_controller.py
│           - Manages computational workloads using SkyPilot
│
└── pipelines/              # Data and ML pipeline components
    ├── data_engineering/   # Data preparation stage
    │   └── nodes.py
    │       - Load raw data
    │       - Clean and preprocess data
    │       - Generate features
    │
    ├── machine_learning/   # Model development stage
    │   └── nodes.py
    │       - Split data into train/test sets
    │       - Train machine learning models
    │       - Evaluate model performance
    │
    └── model_deployment/   # Model deployment and monitoring
        └── nodes.py
            - Package trained models
            - Deploy models to target environments
            - Monitor model performance

Development Setup:

$ python3 -m pip install --upgrade pip
$ python3 -m pip install --user pipx
$ python3 -m pipx ensurepath
$ pipx install poetry==1.8.3
$ cp .env.example .env
$ direnv allow # See https://direnv.net/#getting-started
$ make
$ timestep up --dev

Library Setup:

$ python3 -m pip install --upgrade pip
$ python3 -m pip install --user pipx
$ python3 -m pipx ensurepath
$ pipx install timestep
$ timestep up

Usage:

$ timestep [OPTIONS] COMMAND [ARGS]...

Options:

• --install-completion: Install completion for the current shell.
• --show-completion: Show completion for the current shell, to copy it or customize the installation.
• --help: Show this message and exit.

Commands:

• up: Start up the Timestep AI platform.

timestep up

Start up the Timestep AI platform.

Usage:

$ timestep up [OPTIONS]

Options:

• --allowed-image-ids TEXT: Allowed image IDs to filter by [default: ami-0e7c4f6b17a66658a]
• --allowed-image-names TEXT: Allowed image names to filter by [default: 24.04 (LTS) x64, Ubuntu 24.04 LTS]
• --allowed-location-countries TEXT: Allowed location countries to filter by
• --allowed-location-ids TEXT: Allowed location IDs to filter by
• --allowed-location-names TEXT: Allowed location names to filter by
• --clean / --no-clean: Clean up [default: no-clean]
• --dev / --no-dev: Development mode [default: no-dev]
• --host TEXT: Host [default: 0.0.0.0]
• --min-bandwidth INTEGER: Minimum bandwidth in GB
• --min-disk INTEGER: Minimum disk size in GB [default: 10]
• --min-ram INTEGER: Minimum RAM in MB [default: 2000]
• --name TEXT: Name [default: timestep]
• --port INTEGER: Port [default: 8000]
• --ssh-key TEXT: Path to the SSH key [default: ~/.ssh/id_ed25519]
• --help: Show this message and exit.