Skip to content

Getting Started

Sign in to the Runta dashboard, then create and copy a key from Settings → Runta API Keys.

Terminal window
export RUNTA_TOKEN="rt..."

Runta setup currently targets macOS and Linux. Windows is not supported yet.

Linux:

Terminal window
npm install -g @runta/runta-cli

macOS (Apple silicon):

Terminal window
brew install runta-dev/tap/runta
Terminal window
runta run --name demo --cpus 2 --memory 2048
runta exec demo -- python3 -c 'print("Hello from Runta")'
runta rm demo

Choose either the dashboard or the CLI to create an agent runtime.

Open the Runta dashboard, create a runtime, set its name and compute resources, then choose an agent such as Codex CLI, Claude Code, or OpenClaw.

CREATE RUNTIMEDashboard preview

Create runtime

Set the runtime details, compute resources, and agent setup.

Open dashboard →
Runtime details Name: my-runtimeCPU 1 vCPUMemory 2 GiB
Clean runtime
Codex CLI
Claude Code
OpenClaw

Create an OpenAI API key.

Pass the --agent parameter when creating a runtime. Runta installs and configures the selected coding agent automatically.

Store the exported OpenAI API key in Runta once, then create the runtime:

Terminal window
runta secret set OPENAI_API_KEY --value-env OPENAI_API_KEY
runta run --name codex-demo --cpus 2 --memory 2048 --agent codex

The Codex agent preset automatically installs Codex and injects the stored credential into matching OpenAI requests without exposing the real key inside the runtime.