OpenAI Codex

June exposes a native OpenAI /responses API, which means you can use Codex directly through June.

How It Works

Codex uses the OpenAI responses format under the hood. By pointing it at June's API and providing your June API key, all requests are routed through June's infrastructure. This gives you the same privacy and billing benefits you're used to from June, within a coding agent of your choice.

Setup

Edit .codex/config.toml to include the following section:

[model_providers.june]
name = "June"
base_url = "https://api.blockchain.info/ai/api/v1"
experimental_bearer_token = "<JUNE_API_KEY>"

Replace <JUNE_API_KEY> with your June API key.

Then run codex -c "model_provider=june".

If you want to make this the default, you can add model_provider = "june" to the top of your .codex/config.toml file, e.g.

model = "gpt-5.4"
model_provider = "june"
model_reasoning_effort = "high"
personality = "pragmatic"

[model_providers.june]
name = "June"
base_url = "https://api.blockchain.info/ai/api/v1"
experimental_bearer_token = "<JUNE_API_KEY>"

Model Selection

You can also run Codex using a different model, for example if you want to use blockchain/june, you can run codex -c "model_provider=june" --model "blockchain/june".

Alternatively, you can set the model property of your .codex/config.toml file.

See the models guide for more information on supported models.