Get Started

1. Install InitRunner

The installer sets up the initrunner CLI with everything you need:

2. Setup

Pick your AI provider and paste your API key:

initrunner setup

Supports OpenAI, Anthropic, Google, Groq, Mistral, Cohere, xAI, AWS Bedrock, and Ollama. Re-run any time, or use initrunner configure to switch providers or models for individual agents.

3. Install from the hub

Search for agents from the CLI or browse InitHub, then install with a single command:

initrunner search "code review"

Run any installed agent by name:

initrunner run package-name -p "your prompt"

If an installed agent uses a provider you haven't configured, the CLI prompts you to adapt it. You can also switch later with initrunner configure agent-name --provider NAME or initrunner configure agent-name --model NAME. Use initrunner list to see installed agents and initrunner update --all to keep them current.

4. Create your own

Describe what you want and InitRunner scaffolds a ready-to-run agent:

initrunner new "describe your agent"

Run it interactively to iterate:

initrunner run role.yaml -i

Use --from hub:owner/name to fork a community agent, or --blank for an empty scaffold. The wizard generates reasoning and autonomy specs when you describe planning-style agents.