Codex is the default provider in DP Code. When you start a session with theDocumentation Index
Fetch the complete documentation index at: https://mintlify.com/Emanuele-web04/dpcode/llms.txt
Use this file to discover all available pages before exploring further.
codex provider, DP Code launches codex app-server as a subprocess and communicates with it using JSON-RPC over stdio. All agent output is streamed to the browser through the DP Code WebSocket server.
Prerequisites
Before you can use the Codex provider, you need the Codex CLI installed and authorized on the machine running the DP Code server.Install Codex CLI
Follow the installation instructions in the Codex repository for your platform.
Authorize Codex
Run
codex at least once in a terminal to complete the login flow and confirm your account is active.Settings
Configure the Codex provider underproviders.codex in your DP Code server settings.
Path to the
codex executable. Accepts a bare binary name resolved from PATH or an absolute path. Leave empty to use the default "codex".Override the Codex home directory (equivalent to setting
CODEX_HOME). When set, DP Code uses this path for Codex configuration and state instead of the default location. Leave empty to use Codex’s standard home directory.List of additional model slugs to make available in the model picker alongside the built-in options. Each entry must be a non-empty string with a maximum of 256 characters.
Set to
false to disable the Codex provider entirely. Existing sessions are not affected until you reload.Model selection
The following models are available by default. DP Code selectsgpt-5.5 when no model is specified.
| Model slug | Display name | Default reasoning effort |
|---|---|---|
gpt-5.5 | GPT-5.5 | Medium |
gpt-5.4 | GPT-5.4 | High |
gpt-5.4-mini | GPT-5.4 Mini | High |
gpt-5.3-codex | GPT-5.3 Codex | High |
gpt-5.3-codex-spark | GPT-5.3 Codex Spark | High |
gpt-5.2-codex | GPT-5.2 Codex | High |
gpt-5.2 | GPT-5.2 | High |
customModels if you have access to early-access or private models.
Model aliases
DP Code accepts short aliases in addition to canonical slugs:| Alias | Resolves to |
|---|---|
5.5 | gpt-5.5 |
5.4 | gpt-5.4 |
5.3 | gpt-5.3-codex |
gpt-5.3 | gpt-5.3-codex |
5.3-spark | gpt-5.3-codex-spark |
Reasoning effort options
All Codex models support the following reasoning effort levels. You set this per turn in the model options UI.| Value | Label |
|---|---|
low | Low |
medium | Medium |
high | High |
xhigh | Extra High |
gpt-5.5 defaults to medium. All other models default to high.
Fast mode
Models based on the GPT-5 family support fast mode, which reduces latency at the cost of response depth. Toggle it in the model options panel.Approval policies and sandbox modes
DP Code maps its two runtime modes to Codex approval and sandbox settings automatically.| DP Code runtime mode | Codex approval policy | Codex sandbox |
|---|---|---|
approval-required | untrusted | read-only |
full-access | never | danger-full-access |
How the session works
When you open a thread with the Codex provider, DP Code:- Spawns
codex app-serveras a subprocess with JSON-RPC over stdio. - Sends an
initializerequest to negotiate capabilities. - Opens or resumes a Codex thread with
thread/startorthread/resume. - Forwards each user message as a
turn/startJSON-RPC request. - Streams events back to your browser over the DP Code WebSocket connection.
Codex repository
Source code, releases, and installation instructions for the Codex CLI.
Codex app-server docs
Official documentation for the Codex app-server JSON-RPC protocol.