Runtimes

Choose the software that runs your agent and connect it to a model provider.

Set spec.runtime to choose the agent software and image. Choose its model and credentials in the provider. spec.model is ignored. Without a managed provider, the runtime uses its default model.

Supported providers

Runtime	Provider
`ClaudeCode`	`anthropic`
`Codex`	`openai`
`Gemini`	`gemini`
`Pi`	`anthropic` \| `openai` \| `gemini` \| `groq` — a key is required
`OpenCode`	`anthropic` \| `openai` \| `zhipu` \| … override
`Hermes`	`anthropic` \| `openai` \| `gemini` \| … — a key is required
`Hal`	`anthropic` \| `openai` \| `openai_compatible` \| `openrouter` — a key is required. Industric's own harness.

OpenCode runs keyless

OpenCode can use a free model without a provider. Other runtimes need one.

Examples

Choose an example and set runtime and provider for your setup. Set the model in the provider configuration.

claudecode.yaml· yaml

1apiVersion: autonomic.sh/v1alpha1
2kind: Agent
3metadata:
4  name: claudecode-agent
5spec:
6  runtime: ClaudeCode
7  mode: supervised
8  phase: Active
9  goal:
10    inline: |
11      Implement issues labelled good-first-issue in the billing service.
12      Add a test for every behaviour change and open a draft PR.
13  provider:
14    type: anthropic
15    authMode: apiKey
16  bootstrap:
17    skills:
18      - write-tests

Image versions

Choose image versions on the deployment or resource page. Apply them on the next restart or restart now. These controls are available in the 11 November 2026 build; older releases may differ. See Runtime images for overrides.

Pinned versions

Pinning keeps an image version fixed. Update or remove the pin to receive newer security and compatibility fixes.