Skip to main content
Codex supports custom model providers through config.toml, so you can route Codex requests through Phaseo while keeping the same CLI, app, or IDE workflow. If you also want Codex to read the live Phaseo documentation while it works, add the Docs MCP server as well.
Need reusable agent setup help as well as provider wiring? Install the public Phaseo skills from phaseoteam/skills:
npx skills add phaseoteam/skills --agent codex --skill '*'

​
Prerequisites

  • A Phaseo API key.
  • Codex CLI, Codex App, or the Codex IDE extension.

​
1) Set your API key

export PHASEO_API_KEY="your_phaseo_api_key"

$env:PHASEO_API_KEY = "your_phaseo_api_key"

​
2) Configure Codex to use Phaseo

Create or update your user config at ~/.codex/config.toml. For a trusted project, you can also place the same settings in .codex/config.toml inside the repo. 
model = "openai/gpt-5.5"
model_provider = "phaseo"

[model_providers.phaseo]
name = "Phaseo"
base_url = "https://api.phaseo.ai/v1"
env_key = "PHASEO_API_KEY"
wire_api = "responses"
Codex currently uses the Responses wire protocol for custom providers. Keep wire_api = "responses" or omit it and use the default.

​
3) Start Codex

codex
The Codex App and IDE extension share the same config layers. In the IDE extension, open the config from Codex Settings > Open config.toml.

​
Verification

  • Confirm your model id exists in Models before launching.
  • Use Responses for the default Codex-style flow.

​
Troubleshooting

  • 401 Unauthorized: check that PHASEO_API_KEY is set in the same shell running codex.
  • model_not_found: fetch a current id from Models and update model.
  • Provider-specific issues: follow Routing and fallbacks for failover behavior.
Last modified on July 9, 2026