Runtime for OpenCode

OpenCode support is coming soon. When it lands, it will run as one harness inside Spinup's cloud agent runtime, with scoped credentials per provider so Claude, OpenAI, OpenRouter, Ollama, and open-weight model routes can change without rebuilding the agent. Today, OpenClaw and Hermes already run on this model.

Coming Soon

What OpenCode needs from a cloud runtime

OpenCode's strength is its simplicity: a coding agent that works with many models. Each piece below has to land before Spinup can call it supported.

One environment, any provider

OpenCode works with OpenAI, Anthropic, Google, OpenRouter, Ollama, and many other providers. The work is to project a scoped credential per provider so switching between Claude, OpenAI, Kimi, DeepSeek, GLM, OpenAI OSS models, or the next route is a config change.

Updates that don't churn the environment

OpenCode is open-source, maintained by Anomaly with community contributions. The harness lifecycle has to tolerate upstream updates without destabilizing runs.

State that survives between runs

Installed packages, generated code, project state, and tool configuration each need a clear persistence contract so runs can resume.

Unattended runs

Moving from terminal to cloud means running unattended, on a schedule, or as part of a team workflow. That contract gets defined when OpenCode lands.

Local to Cloud

The gap between terminal agent and cloud runtime

OpenCode works well in the terminal. You install it, configure your model provider, and run it against a local codebase. The agent reads files, writes code, runs commands, and iterates. Execution happens on your machine, in your shell, with your environment variables.
Cloud runs change the picture. API keys cannot live in your shell profile. The filesystem needs to persist across sessions without depending on your local disk. Network access requires policy-level controls, not just whatever your machine allows. The agent needs to run reliably without a terminal window open.
You can't run OpenCode on Spinup yet. OpenClaw and Hermes are the supported harnesses today, on the same isolation, secret, and persistence model. OpenCode plugs into the same shape once its specifics are ready.

FAQ

Common questions about OpenCode on Spinup

How does Spinup handle OpenCode's multi-provider model support?+

Not yet. OpenCode is on the roadmap. When it lands, provider keys for OpenAI, Anthropic, Google, OpenRouter, Ollama, and open-weight routes such as Kimi, DeepSeek, GLM, or OpenAI OSS models get projected as scoped secrets where the provider supports them. Switching providers should be a config change rather than a rebuild.

What does a cloud runtime add to a terminal-based agent?+

Persistent environments, managed secrets, network policy, and unattended runs: the parts a terminal session does not give you. OpenCode is not a supported harness yet, so today the cloud runtime ships with OpenClaw and Hermes.

Can I compare OpenCode with other harnesses in the same environment?+

Not yet. Today the supported swap is between OpenClaw and Hermes. OpenCode joins the same portable runtime model when it lands.

How does the open-source update model affect the runtime?+

OpenCode is open-source and maintained by Anomaly with community contributions, so updates land on a community schedule. The harness lifecycle has to absorb that pace without letting upstream drift turn into runtime instability. That contract is part of the OpenCode work that comes after launch.

Related

The runtime should stay bigger than any one harness

Related pages

What is a Cloud Agent Runtime?→Secure Code Execution for AI Agents→Runtime for Codex CLI→Runtime for Hermes→

Early access

OpenCode is on the way. Try OpenClaw or Hermes on the same runtime today.

Request access if this is the runtime shape your team has been missing.

We reply with a booking link and review the fit before inviting your workspace.