Prime Intellect CLI & SDKs

# Install uv (if not already installed)
curl -LsSf https://astral.sh/uv/install.sh | sh
Install prime
uv tool install prime
Authenticate
prime login
Set up a Lab workspace for environments, evals, GEPA, and Hosted Training
prime lab setup
See available Hosted Training models, capacity, and pricing
prime train models
Launch a prime-rl config as a Hosted Training run
prime train rl.toml
Browse verified environments
prime env list
List available GPU resources
prime availability list

curl -LsSf https://astral.sh/uv/install.sh | sh

uv tool install prime

pip install prime

uv pip install prime-sandboxes

See prime-sandboxes documentation for SDK usage.

Usage

Configuration

API Key Setup

# Interactive mode (recommended - hides input)
prime config set-api-key
Non-interactive mode (for automation)
prime config set-api-key YOUR_API_KEY
Environment variable (most secure for scripts)
export PRIME_API_KEY="your-api-key-here"

Other Configuration

# Configure SSH key for pod access
prime config set-ssh-key-path
View current configuration
prime config view

Security Note: When using non-interactive mode, the API key may be visible in your shell history. For enhanced security, use interactive mode or environment variables.

Environments Hub

Access hundreds of verified environments on our community hub with deep integrations with sandboxes, training, and evaluation stack.

# Browse available environments
prime env list
View environment details
prime env info <environment-name>
Inspect environment source without downloading the archive
prime env inspect <environment-name>
Install an environment locally
prime env install <environment-name>
Push your own environment (scaffold one with verifiers' vf-init)
prime env push my-environment

Lab and Hosted Training

Start with prime lab setup to create a local workspace with starter configs, coding-agent skills and a verifiers install; build and evaluate environments there with verifiers' vf-init and vf-eval. Then use prime train models to choose a Hosted Training model with current capacity and pricing. Hosted Training runs take prime-rl configs as-is; its examples/ are good starting points.

# Set up a Lab workspace
prime lab setup
List cached models and on-demand GPU capacity and pricing
prime train models
Launch a prime-rl config (RL, or SFT with --sft) on a dedicated cluster
prime train rl.toml
Inspect and manage Hosted Training runs
prime train list
prime train logs <run-id> -f
prime train metrics <run-id>
prime train checkpoints <run-id>

GPU Resources

# List all available GPUs
prime availability list
Filter by GPU type
prime availability list --gpu-type H100_80GB
Show available GPU types
prime availability gpu-types

Pod Management

# List your pods
prime pods list
Create a pod
prime pods create
prime pods create --id <ID>     # With specific GPU config
prime pods create --name my-pod # With custom name
Monitor and manage pods
prime pods status <pod-id>
prime pods terminate <pod-id>
prime pods ssh <pod-id>

Evaluations

Push and manage evaluation results to the Environments Hub.

# Auto-discover and push evaluations from current directory
prime eval push
Push specific eval directory (verifiers format)
prime eval push outputs/evals/gsm8k--gpt-4/abc123
Push a public evaluation (default is private)
prime eval push --public
List all evaluations
prime eval list
Get evaluation details
prime eval get <eval-id>
View evaluation samples
prime eval samples <eval-id>

Team Management

# List teams
prime teams list
Switch context
prime switch
prime switch personal
prime switch <team-slug>

Development

# Clone the repository
git clone https://github.com/PrimeIntellect-ai/prime
cd prime
Set up workspace (installs all packages in editable mode)
uv sync
Install CLI globally in editable mode
uv tool install -e packages/prime
Now you can use the CLI directly
prime --help
Run tests
uv run pytest packages/prime/tests
uv run pytest packages/prime-sandboxes/tests

All workspace packages (prime, prime-sandboxes, prime-evals, prime-tunnel) are installed in editable mode. Changes to code are immediately reflected.

Releasing

This monorepo contains independently versioned packages: prime (CLI + full SDK), prime-sandboxes, prime-evals, and prime-tunnel (lightweight SDKs).

Versions are single-sourced from each package's __init__.py file:

• prime: packages/prime/src/prime_cli/__init__.py
• prime-sandboxes: packages/prime-sandboxes/src/prime_sandboxes/__init__.py
• prime-evals: packages/prime-evals/src/prime_evals/__init__.py
• prime-tunnel: packages/prime-tunnel/src/prime_tunnel/__init__.py

To release a new version:

1. Update the __version__ string in the appropriate __init__.py file
2. Commit and push the change

Version sync considerations:

When releasing prime, consider whether prime-sandboxes or prime-tunnel should also be bumped, as prime depends on both. The packages can be released independently or together depending on what changed.

License

This project is licensed under the MIT License - see the LICENSE file for details.

Links

• Website
• Dashboard
• API Docs
• Discord Community