Set up Codex

01 / PREPARE

Install the agent and prepare access

1. Install the agent using its official guide. Check its version before merging the examples below.
2. Purchase your DeepSeek plan, then open Billing & API keys to copy your active key. OpenAI base: https://ds-api.savemytoken.com/v1. Claude Code base: https://ds-api.savemytoken.com. DeepSeek credentials are a separate scope.
3. Back up the file you will edit. Merge fields into existing objects and tables; preserve unrelated providers, tools and permission settings. Keep credentials in private local settings.
4. Make SAVEMYTOKEN_API_KEY available to the terminal launching Codex. A custom CODEX_HOME changes the configuration directory; use that directory consistently.

Ready when: the agent opens, and you know which plan, process and configuration directory you will use.

02 / CONFIGURE

Save the connection and model catalog

Create the parent directories if needed. Download or copy each file, replace the example endpoint and local placeholders, then merge it into the stated location. Downloads and the examples below contain the same configuration.

model_provider = "savemytoken-deepseek"
model = "deepseek-flash"
model_catalog_json = "/ABSOLUTE/PATH/TO/savemytoken-deepseek-models.json"

[model_providers.savemytoken-deepseek]
name = "SaveMyToken"
base_url = "https://ds-api.savemytoken.com/v1"
env_key = "SAVEMYTOKEN_API_KEY"
wire_api = "responses"
requires_openai_auth = false
supports_websockets = false

{
  "models": [
    {
      "slug": "deepseek-flash",
      "display_name": "DeepSeek Flash",
      "description": "SaveMyToken DeepSeek route",
      "default_reasoning_level": "high",
      "supported_reasoning_levels": [
        {
          "effort": "none",
          "description": "Off"
        },
        {
          "effort": "low",
          "description": "Low"
        },
        {
          "effort": "high",
          "description": "High"
        },
        {
          "effort": "max",
          "description": "Max"
        }
      ],
      "shell_type": "unified_exec",
      "visibility": "list",
      "supported_in_api": true,
      "priority": 0,
      "base_instructions": "You are a coding assistant. Follow the user's instructions, inspect relevant files, preserve existing work, and verify changes. Respect configured permissions and approvals.",
      "supports_reasoning_summaries": false,
      "supports_reasoning_summary_parameter": false,
      "support_verbosity": false,
      "default_verbosity": null,
      "apply_patch_tool_type": null,
      "truncation_policy": {
        "mode": "bytes",
        "limit": 10000
      },
      "experimental_supported_tools": [],
      "input_modalities": [
        "text"
      ]
    }
  ]
}

Ready when: the files parse successfully and the selected profile points to the delivered URL, key reference and exact model catalog.

03 / START

Start the agent and choose your model

Open a terminal in a trusted project. Complete the file setup before running these commands.

export SAVEMYTOKEN_API_KEY="REPLACE_WITH_YOUR_DELIVERED_KEY"
codex --profile savemytoken-deepseek

Run /model and select a model from this catalog, then choose its effort. /status shows the current session. Keep the SaveMyToken provider selected; a model name alone does not choose a gateway.

Check: the provider is the one you configured and the selected model is deepseek-flash. A visible model list alone does not prove inference works.

04 / REASONING

Choose and verify reasoning effort

Each catalog entry supplies supported_reasoning_levels and a non-null default_reasoning_level. Remove an old profile-level model_reasoning_effort if you want the model defaults.

A command-line override or saved task can supersede the default. After switching models, check the effort again. Do not copy one universal high/low setting onto models with different controls.

/model
/status

On DeepSeek Flash, choose Low, then High, then Low again. Inspect the active setting each time before sending the next message. A longer answer is not evidence that effort changed.

REASONING CONTROLS

Choose effort for the actual model

The setup templates start with DeepSeek Flash. Add other delivered routes only after confirming their exact IDs and capabilities.

05 / DESKTOP

Codex Desktop · local tasks

Desktop reads the main config.toml. Choose the macOS or Windows example below, and merge its top-level provider/model/catalog settings before any TOML tables. CLI profiles are not a substitute for these Desktop defaults.

Save your delivered key as a private local savemytoken.key file containing only the key. Replace the absolute key-file and catalog paths. On macOS, restrict the file to your account (chmod 600); on Windows, restrict access in the file's Security properties. The command-auth provider reads this file even when the app is opened outside a terminal.

Fully quit and reopen Codex, then start a new local task. Check both model and effort. Desktop versions can filter catalog levels; if your required level is missing, use CLI. Neither the catalog nor a visible picker proves a successful model request.

model_provider = "savemytoken-deepseek"
model = "deepseek-flash"
model_catalog_json = "/ABSOLUTE/PATH/TO/savemytoken-deepseek-models.json"

[model_providers.savemytoken-deepseek]
name = "SaveMyToken"
base_url = "https://ds-api.savemytoken.com/v1"
wire_api = "responses"
supports_websockets = false

[model_providers.savemytoken-deepseek.auth]
command = "/bin/cat"
args = ["/ABSOLUTE/PATH/TO/savemytoken.key"]

model_provider = "savemytoken-deepseek"
model = "deepseek-flash"
model_catalog_json = "C:/ABSOLUTE/PATH/TO/savemytoken-deepseek-models.json"

[model_providers.savemytoken-deepseek]
name = "SaveMyToken"
base_url = "https://ds-api.savemytoken.com/v1"
wire_api = "responses"
supports_websockets = false

[model_providers.savemytoken-deepseek.auth]
command = 'C:\Windows\System32\WindowsPowerShell\v1.0\powershell.exe'
args = ['-NoProfile', '-NonInteractive', '-Command', 'Get-Content -Raw -LiteralPath "C:\ABSOLUTE\PATH\TO\savemytoken.key"']