Codex is OpenAI's coding agent, available as a command-line tool and a VS Code extension. With a custom provider, Codex can use models on DataEyesAI.Before you begin#
| Setting | Value |
|---|
| Base URL | https://studio.dataeyes.ai/v1 (/v1 required) |
| API key | Your Model API key, passed via the DATAEYES_API_KEY environment variable |
| Recommended models | gpt-6-astra, gpt-6-luna, gpt-6.1-sol |
Codex uses the OpenAI Responses protocol (/v1/responses). Choose a GPT-6 series model.
1. Install Codex#
Requires Node.js 18 or later. Run codex --version to confirm the installation.2. Write the config file#
Open or create ~/.codex/config.toml (on Windows: C:\Users\<username>\.codex\config.toml) and add:model_provider = "dataeyes"
model = "gpt-6-astra"
preferred_auth_method = "apikey"
disable_response_storage = true
[model_providers.dataeyes]
name = "DataEyesAI"
base_url = "https://studio.dataeyes.ai/v1"
env_key = "DATAEYES_API_KEY"
wire_api = "responses"
| Field | Description |
|---|
model_provider | Points to the dataeyes provider defined below |
model | Default model; switch in a session with /model |
base_url | DataEyesAI address; must end with /v1 |
env_key | Environment variable that holds the API key |
wire_api | Always responses |
3. Set the API key#
To make it permanent, add it to ~/.zshrc or ~/.bashrc. 4. Verify#
Run codex in any project directory and ask "Summarize the structure of the current directory". A normal response means the setup works. You can also run a quick non-interactive test:Use in VS Code#
The Codex VS Code extension shares ~/.codex/config.toml and environment variables with the CLI. Once the steps above are done, install the Codex extension from the marketplace and use it directly.FAQ#