Codex

Use Bitdeer AI models in OpenAI's Codex app.

Codex is OpenAI's AI coding agent. By adding Bitdeer AI as a custom model provider, you can route Codex requests to Bitdeer AI models such as GLM-5.3 — no OpenAI account required.

Prerequisites

Configuration steps

Sign out of Codex. If you're currently signed in to Codex with a ChatGPT account, sign out before continuing.

Open the Codex configuration file in a text editor. Create it if it doesn't exist. (~/.codex/config.toml for macOS/Linux or %USERPROFILE%\.codex\config.toml for Windows)

Add the following configuration, replacing BITDEER_API_KEY with your Bitdeer AI API key:

model_provider = "bitdeer"
model = "zai-org/GLM-5.3"
model_reasoning_effort = "high"
disable_response_storage = true

[model_providers.bitdeer]
name = "Bitdeer AI"
base_url = "https://api-inference.bitdeer.ai/v1"
wire_api = "responses"
requires_openai_auth = false
experimental_bearer_token = "BITDEER_API_KEY" # Replace with your API key

Important

  • The first four lines must appear above any line that starts with [ in the file.
  • If any of these keys already exist in your file, replace them instead of adding duplicates. Duplicate keys will prevent Codex from starting.
  • Your API key is stored in plain text. Keep this file private and never share or commit it. On macOS / Linux, you can restrict access with chmod 600 ~/.codex/config.toml.

Restart Codex. Fully quit and reopen Codex. The model selector should now display custom. Send a test prompt to confirm everything is working.

Notes

  • To switch models, change the model value in config.toml and restart Codex.
  • To use Codex with your ChatGPT account again, remove the lines added above from config.toml, restart Codex, and sign in.

For list of all available models, checkout Model Catalog

On this page