Does Privatemode store my source code or prompts?

No. All inference happens inside hardware-isolated confidential computing environments. Your prompts and code context are encrypted before leaving your machine, decrypted only inside the secure enclave for processing, and discarded after the response is generated. Nothing is stored or used for model training.

How does encryption affect response speed?

The Privatemode proxy adds only milliseconds for encryption, decryption, and remote attestation. Response times are comparable to standard cloud AI APIs. You can also enable prompt caching with the --sharedPromptCache flag on the proxy to speed up repeated context windows.

Can I verify that the server is actually running in a secure enclave?

Yes. The Privatemode proxy performs remote attestation before every session, verifying the server's integrity against known cryptographic hashes. The proxy and server code are open source with reproducible builds, so you can audit exactly what runs on the server side.

Does Claude Code work the same way through Privatemode?

Claude Code's core workflow is fully supported: multi-file editing, codebase-wide search, shell command execution, and project reasoning. The Privatemode API is Anthropic-compatible, so Claude Code communicates with it natively. The only difference is the model: instead of Anthropic's Claude models, you use Privatemode's encrypted open-source models.

Using Confidential AI with Claude Code

Claude Code gives developers an AI coding assistant right in the terminal. With Privatemode, every prompt and code snippet stays encrypted end-to-end, so your proprietary codebase never leaves your control.

Introduction

Claude Code integration

Benefits

Why use Claude Code with Privatemode

Introduction

Use Claude Code for agentic coding without exposing your source code

AI coding assistants need access to your codebase

Claude Code is Anthropic's agentic coding tool that operates directly in your terminal. It reads, edits, and reasons across your entire codebase. By default, every prompt and code context is sent to Anthropic's cloud API. For teams working on proprietary software, regulated applications, or sensitive IP, that means source code leaves your control.

Privatemode keeps your code confidential

Privatemode provides an Anthropic-compatible Messages API backed by state-of-the-art models running inside confidential computing environments. Point Claude Code at the Privatemode proxy and your source code, prompts, and AI responses are encrypted end-to-end. Not even Privatemode can see your data.

Hardware-enforced, not just policy

The Privatemode proxy runs locally on your machine and encrypts all data before it leaves. On the server side, inference runs inside hardware-isolated confidential computing environments (AMD SEV / Intel TDX and Nvidia Confidential Computing). The proxy verifies server integrity through remote attestation before every session. Your code is never stored and never used for model training.

Benefits

Why use Privatemode AI with Claude Code?

End-to-end encrypted coding sessions

Every prompt you type and every file Claude Code reads is encrypted by the local Privatemode proxy before leaving your machine. Responses are decrypted only on your device. The models run inside confidential computing environments protected by AMD SEV and Intel TDX hardware. Your source code, architecture decisions, and business logic remain confidential throughout the entire session.

Full agentic coding capabilities

Claude Code's core features work through Privatemode: multi-file editing, codebase search, shell command execution, and project-wide reasoning. Use the gpt-oss-120b model for complex tasks that require deep understanding of your repository, or Qwen3-Coder for fast completions and edits.

Ready in minutes

Set three environment variables in your Claude Code settings to route all traffic through Privatemode. Your existing Claude Code setup gains end-to-end encryption immediately.

Read the docs

Set up Claude Code to use Privatemode

Follow the step-by-step guide in the documentation to set up Claude Code to use Privatemode as inference provider.