Integrate with Codex

Replace the placeholder values in the following snippet with your actual credentials and parameters, then run the script to integrate with F5 AI Security.

Add this snippet to the Codex configuration file, typically ~/.codex/config.toml:`:

toml

model_provider = "f5-ai-security"

[model_providers.f5-ai-security]
name = "F5-AI-security"
wire_api = "responses"
supports_websockets = false
env_key = "CALYPSOAI_TOKEN"
base_url = "{BASE_URL}/openai/{CONNECTION_NAME}"

Ensure the model_provider setting is not nested under another section in the config file, it should be at the top level.

With the following commands:

Set the CALYPSOAI_TOKEN environment variable to your API key
Run codex exec with a question to confirm the connection.

shell

export CALYPSOAI_TOKEN={CAI_TOKEN}
codex exec "What is the capital of France?"

You can integrate F5 AI Security. We provide a code snippet to help with the process, which requires the following variables:

BASE_URL: The URL of the AI Security deployment. Depends on whether your deployment uses AI Security SaaS or your own on-premises system.
- SaaS: The URLs may be:
  - https://us1.calypsoai.app
  - https://eu1.calypsoai.app
- On-premises: URL based on your deployment.
CAI_TOKEN: See Get Your Authorization Token for instructions on how to obtain an API key.
CONNECTION_NAME: Name of the configured OpenAI connection that Codex should use.

Integrate with Codex ​

Integrate with Codex