Codex CLI is OpenAI's open-source coding agent for the terminal. It supports custom model providers through its config file, so you can run it against DEVSAIO.
1. Set your key as an environment variable
export DEVSAIO_API_KEY=YOUR_KEYOn Windows PowerShell: $env:DEVSAIO_API_KEY = "YOUR_KEY".
2. Edit ~/.codex/config.toml
On Windows the file is at %USERPROFILE%\.codex\config.toml.
model = "gpt-5.6-luna"
model_provider = "devsaio"
[model_providers.devsaio]
name = "DEVSAIO"
base_url = "https://devsaio.lol/v1"
env_key = "DEVSAIO_API_KEY"
wire_api = "responses"env_keytells Codex which environment variable holds the key, so the key never sits in the file.wire_api = "responses"uses the Responses API format, which DEVSAIO supports.
3. Run it
codexAsk it something small first (“list the main folders in this repo and what they do”), then check Usage Logs to confirm requests are arriving.
Using other models
Change model to any model from the model list. If a model misbehaves with the Responses format, try wire_api = "chat" instead.
Tips
- Set a monthly limit on the key Codex uses.
- Keep tasks focused; agents spend most tokens reading files.
- The integrations page generates this config with your key name filled in.
Claude, GPT, Gemini, DeepSeek, image and video models through one OpenAI- and Anthropic-compatible endpoint. Prepaid, pay only for what you use.