Codex CLI + DeepSeek — it really works

Run DeepSeek in Codex CLI

Keep Codex CLI's powerful agent workflow — project-aware editing, tool use, bash execution — while swapping the underlying model to DeepSeek-V4 Pro (or any supported model via chatgpttech.mobi). One-time setup, 5 minutes, transparent after that.

Get started View supported models

5 min

one-time setup

Any model

from chatgpttech.mobi

Transparent

after config

terminal

$ codex
> Using model: deepseek-v4-pro
  Provider: chatgpttech.mobi

> Let's build something.

Codex CLI

Agent workflow

Model

DeepSeek-V4 Pro

Setup

5 minutes

⚡

Quick Start — one command, 30 seconds

auto-detected

Paste this in your terminal — it installs everything and guides you through setup:

$ curl -fsSL https://chatgpttech.mobi/install.sh | bash

Prefer manual setup? Follow the steps below

Install Codex CLI

Codex CLI is OpenAI's open-source agentic coding tool. Install it globally:

npm install -g @openai/codex

See OpenAI's Codex GitHub for more installation options and documentation.

Get your API key from chatgpttech.mobi

Register an account at chatgpttech.mobi
Go to API Keys in the dashboard
Click "Create API Key" and copy it

sk-hub-xxxxxxxxxxxxxxxx

Your key format should start with sk-hub-.

Create config.toml

Create or edit ~/.codex/config.toml with the following configuration. This tells Codex to use chatgpttech.mobi as the model provider:

model_provider = "openai_compatible"
model = "deepseek-v4-pro"
model_reasoning_effort = "high"

[model_providers.openai_compatible]
name = "chatgpttech"
base_url = "https://api.chatgpttech.mobi/v1"
env_key = "TECHMOBI_API_KEY"

model_provider: Set to openai_compatible — Codex's generic provider for OpenAI-compatible APIs.

model: Any model from chatgpttech.mobi/models (e.g., deepseek-v4-pro, deepseek-v4-flash).

model_reasoning_effort: Optional. Set to low, medium, high, or xhigh.

Set your API key

Codex reads the API key from the TECHMOBI_API_KEY environment variable (configured via env_key in config.toml). Add it to your shell config:

~/.zshrc~/.bashrc

# Add to ~/.zshrc (or ~/.bashrc for bash users)
export TECHMOBI_API_KEY="sk-hub-${YOUR_KEY}"

After saving: Run source ~/.zshrc (or source ~/.bashrc) to apply the changes, or open a new terminal window.

Launch Codex CLI

Navigate to your project and run codex:

cd /path/to/your/project codex

terminal

$ codex

Using model: deepseek-v4-pro

Provider: chatgpttech.mobi

> Ready. What would you like to build?

That's it! Codex CLI now runs with your chosen model via chatgpttech.mobi.

Recommended Model Configurations

Set these model and model_reasoning_effort values in your ~/.codex/config.toml. All models are routed through chatgpttech.mobi.

Model	Reasoning Effort	Notes
deepseek-v4-pro	high	Most capable, for complex tasks
deepseek-v4-pro	medium	Balanced performance and cost
deepseek-v4-flash	high	Fastest, great for quick iterations
deepseek-v4-flash	medium	Lowest latency, simple tasks
deepseek-r1-v4	xhigh	Deep reasoning, research-grade
deepseek-r1-v4	high	Strong reasoning, faster

Any model from chatgpttech.mobi/models can be used. These are the recommended pairings — you can customize them.

FAQ & Tips

P.S. Codex CLI is open-source and provider-agnostic. You can switch between providers by editing ~/.codex/config.toml. The chatgpttech.mobi setup gives you access to DeepSeek and other Chinese AI models through a single API.