Articles

OpenCode: what it is and how we use it

An open agent that connects to a model running on your side as well as to cloud providers.

1. How it works

OpenCode is an agent for working with code and other tasks that runs in the terminal. It is open and can connect to models from various providers, including ones running on your own hardware.

2. Why it matters to a business

For a business it means the tool can be used with the model you pick according to the sensitivity of your data and your budget. You are not tied to one supplier.

3. How we use it

A harness is a coding or assistant tool that connects to an AI model. We set it up and connect it to a model to fit your brief.

4. When it fits

It fits when you want one tool for models on your side and in the cloud.

5. What to watch for

The quality of the result depends on the model chosen.

6. In practice

We deploy OpenCode as a working tool for development and automation. We choose the model per task: local where data must not leave the company, or cloud where strength matters. We set access and permissions so the tool does only what it may.

7. Alternatives

Similar tools are Claude Code, Codex or Cursor, which are mostly tied to one provider. OpenCode is open and the model can be swapped. Which tool fits we verify on a real task.

8. Common questions

Does it work with a local model?

Yes, it can be connected to a model running on your hardware. The quality of the result then depends on the chosen model.

Can the tool change files by itself?

Depending on settings. We recommend a mode where it proposes changes first and you approve them.

Is it free?

The tool itself is open. You pay for the model it uses if it is a cloud provider.

If you need help with this, take a look at our AI infrastructure.

Related articles

• Hermes: what it is and how we use it
• OpenClaw: what it is and how we use it
• BERCIK Code: what it is and how we use it
• AI infrastructure