Use Confidential AI with OpenCode
OpenCode is an open-source coding agent for the terminal. It reads your project, edits files, and runs commands, and it lets you choose which AI provider answers. With Privatemode as that provider, your code goes through a small proxy on your machine to open-weight models running inside confidential computing, where it stays encrypted in memory.
AI-assisted coding shouldn't mean exposing your code
The challenge
Your codebase leaves with every prompt
OpenCode reads across your project to generate and edit code, and sends that context to a cloud LLM. With a conventional provider, your source, architecture decisions, and business logic are processed in plaintext on the provider's servers.
The solution
A provider that cannot read your code
The Privatemode proxy runs locally. It verifies the integrity of the Privatemode cloud service through remote attestation, and encrypts every request before it leaves your machine. Cloud-side, your data is protected end-to-end by confidential computing and stays encrypted in memory even during processing. By design, neither the infrastructure provider nor we can read it. OpenCode lists Privatemode as a provider, so once the proxy runs you select it and pick a model.
How to set up Privatemode in OpenCode
OpenCode never talks to the Privatemode cloud directly. It talks to a proxy on your own machine, and that proxy is what verifies the service through remote attestation and encrypts every request before it leaves. So the proxy needs to be reachable from wherever you use OpenCode.
Frequently asked questions about OpenCode with Privatemode
Privatemode offers a range of models running inside confidential computing environments, with your data encrypted in memory during processing. See the model overview.