𝗥𝘂𝗻 𝗖𝗼𝗱𝗲𝘅 𝗖𝗟𝗜 𝘄𝗶𝘁𝗵 𝗚𝗲𝗺𝗺𝗮 𝟰 𝗼𝗻 𝗪𝗦𝗟𝟮

𝗥𝘂𝗻 𝗖𝗼𝗱𝗲𝘅 𝗖𝗟𝗜 𝘄𝗶𝘁𝗵 𝗚𝗲𝗺𝗺𝗮 𝟰 𝗼𝗻 𝗪𝗦𝗟𝟮

You want a local AI for coding. Run Codex CLI with Gemma 4 on WSL2.

You need these tools:

• llama.cpp
• Node.js
• NVIDIA GPU

Follow these steps:

1. Install Codex CLI. Use npm or curl.
2. Run the codex command. Press Ctrl + C. This creates the .codex folder.
3. Edit ~/.codex/config.toml. Add the llama.cpp provider and port 8080.
4. Start the llama.cpp server. Set the context size to 10000. Small sizes cause errors.
5. Launch Codex. Use the llama provider flag and search flag.

This setup builds React and TypeScript apps locally.

Source: https://dev.to/0xkoji/run-codex-cli-with-local-llm-gemma4-with-llamacpp-on-wsl2-pee Optional learning community: https://t.me/GyaanSetuAi