Codex CLI

Connect OpenAI Codex CLI to Mixen — pay from your balance in roubles, with tool calling and streaming over our OpenAI-compatible /v1/responses.

Installation

npm i -g @openai/codex

Node.js 22+ is required.

Configuration

Create ~/.codex/config.toml (Windows: %USERPROFILE%\.codex\config.toml):

model_provider = "mixen"
model = "openai/gpt-5.5"
model_reasoning_effort = "high"

[model_providers.mixen]
name = "Mixen AI"
base_url = "https://api.mixen.ai/v1/"
env_key = "MIXEN_API_KEY"
wire_api = "responses"

Key points:

• model_provider must match the [model_providers.<id>] section name.
• base_url must end with /v1/.
• wire_api = "responses" — Mixen supports the Responses API (stateless: Codex sends the full history in input — previous_response_id/store are not needed).

The API key environment variable:

export MIXEN_API_KEY="your-api-key"   # add to ~/.bashrc or ~/.zshrc to persist

Running

codex "your prompt"

Recommended models

Model	ID
GPT-5.6 Sol	openai/gpt-5.6-sol
Claude Opus 5	anthropic/claude-opus-5
GLM 5.3	z-ai/glm-5.3
DeepSeek V4 Pro	deepseek/deepseek-v4-pro
Kimi K3	moonshotai/kimi-k3

Full list — in the catalog.

Troubleshooting

• “Model provider not found” — the model_provider name doesn’t match the [model_providers.<id>] section; they must be identical.
• “stream disconnected before completion” — check wire_api = "responses" and the /v1/ suffix in base_url.
• Key not found — the variable from env_key must be set in the current terminal session.
• Model doesn’t call tools — file and command support depends on tool calling in the specific model; try another one from the catalog.