DeepSeek in VS Code Codex: setup and switching FAQ

Question 1

How do I use DeepSeek in the official Codex VS Code extension?

Accepted Answer

Install Codex DeepSeek Switcher (extension ID dzr.codex-local-model-switcher) alongside the OpenAI Codex extension (openai.chatgpt) in the same extension host. Initialize your normal Codex configuration, then run Codex Source: Use DeepSeek Direct. Enter your own DeepSeek API key in the masked prompt when requested, save your work and reload the window. This is an independent companion extension, not an official OpenAI or DeepSeek product.

Question 2

Can I keep my GPT / ChatGPT-authenticated Codex setup while using DeepSeek?

Accepted Answer

The switcher selects the normal .codex home for OpenAI and .codex-vscode-deepseek for DeepSeek. It does not copy the normal auth.json into the DeepSeek home. Windows using the same OS account and home share one provider choice; this is not per-window isolation. Running processes keep their original environment until restarted. Custom CODEX_HOME layouts are not preserved by these launchers.

Question 3

Does using DeepSeek with Codex require MoonBridge or a local proxy?

Accepted Answer

This extension's direct mode uses https://api.deepseek.com/ with wire_api = responses; no MoonBridge or localhost relay is required. It selects provider profiles rather than translating API protocols. It does not supply a DeepSeek account, API credits, model permissions or network connectivity. A ChatGPT login is not used to authenticate the DeepSeek profile.

Question 4

Where do I install the switcher for VS Code Remote SSH?

Accepted Answer

Connect to the Linux SSH host and install both openai.chatgpt and dzr.codex-local-model-switcher under SSH: your-host, not only under Local. The remote side needs its own Codex configuration, Bash, Python 3 and GNU utilities. A separate remote home has a separate provider choice; two SSH aliases or NFS mounts sharing one home do not create isolation.

Question 5

How do I switch from DeepSeek back to OpenAI / GPT?

Accepted Answer

Run Codex Source: Use GPT (OpenAI), save work and reload. This selects the original .codex directory; its existing configuration determines the actual provider. Switching back does not repair a broken original configuration, erase stored DeepSeek credentials or restore the pre-extension executable override. Reapply Current Codex Source Setup is a repair command, not an uninstall command.

Question 6

Is this a universal Codex model switcher for Ollama, Qwen or Claude?

Accepted Answer

No. This implementation targets OpenAI and DeepSeek's native Responses API, with a bundled deepseek-v4-pro / deepseek-v4-flash catalog. Implemented environments are Windows x64 local VS Code and Linux Remote SSH. Local macOS/Linux, WSL, Dev Containers, Codespaces and browser VS Code are not supported. Do not infer support for another provider from an OpenAI-compatible API label.

Question 7

Why does the status say direct: true when my DeepSeek request still fails?

Accepted Answer

Show Current Codex Source Status reports configuration, not live API health. direct: true means the configured endpoint and wire API match direct mode. Test a minimal request in a disposable folder separately, checking network access, model entitlement and account balance; the request may incur charges. Never publish keys or complete configs. The isolated config stores the key in plaintext: masked input is not encryption.