AgentsAgents

Codex

Run Codex from Proliferate against connected repositories and reviewable branches.

Codex can be used as a Proliferate coding agent for implementation, review, and workflow runs, running through OpenAI's own Codex adapter.

Use it when your team wants OpenAI-backed coding workflows inside the same governed workspace model as every other agent.

Supported functionality

CapabilitySupport
Repository workspacesAttach Codex runs to a local checkout, a worktree, or a cloud sandbox with full repo and branch context.
Implementation workRun scoped coding tasks from prompts, issues, or workflows.
Fast modeToggle Codex's fast service tier from the composer for lower-latency runs when you don't need its deepest reasoning.
ReviewInspect generated diffs, the transcript, and command output before merge.
CredentialsSign in with your Codex account, bring an OpenAI API key, or route through the Proliferate Gateway.

Authenticate Codex

Codex authenticates separately for Local runs (on your machine) and Cloud runs (in a Proliferate sandbox). From Settings -> Agents -> Codex you can:

  • Sign in with your Codex account (codex login), which reuses your existing ChatGPT or Codex login state.
  • Add your own OpenAI API key as a BYOK (bring your own key) row.
  • Turn on the Proliferate Gateway for managed access with no setup at all.

Settings -> Agents -> Codex

The Codex harness settings page: Cloud/Local toggle, Authentication tab with the 'Log in with Codex' button, an OpenAI API key row, and the Proliferate Gateway switch.

Chat composer: Fast Mode toggle

The composer control bar for a Codex session showing the Fast Mode toggle alongside the model selector.

For advanced routing, see advanced agent settings.

On this page