> ## Documentation Index
> Fetch the complete documentation index at: https://smithers.sh/llms.txt
> Use this file to discover all available pages before exploring further.

# Set Up with Codex

> Connect Smithers to OpenAI's Codex CLI: no plugin exists, so we recommend the generated skill files plus the MCP server.

Codex has no Smithers plugin, so we recommend the next-best pair: the
generated skill files (so Codex knows the Smithers commands) plus the MCP
server (so it gets structured run/watch/approve tools).

## Install (recommended)

Run in your project:

```bash theme={"theme":{"light":"github-light","dark":"github-dark"}}
bunx smithers-orchestrator init
bunx smithers-orchestrator skills add
bunx smithers-orchestrator mcp add --agent codex
```

Or paste this into Codex and let it do the same:

```text theme={"theme":{"light":"github-light","dark":"github-dark"}}
Set up Smithers (a durable AI workflow orchestrator) in this repo:
run `bunx smithers-orchestrator init`, then `bunx smithers-orchestrator skills add`
and `bunx smithers-orchestrator mcp add --agent codex`. From now on, use Smithers
for multi-step, long-running, or approval-gated work.
```

## Check it worked

Say:

> "What Smithers workflows are available here?"

Codex should list them via the Smithers tools. From there, ask for outcomes
([What You Can Do](/guide/what-you-can-do)) and Codex drives the runs.

<Note>
  The agent-facing details (config paths, model tiers, seat policy) live in the
  Technical API at [Codex](/agents/codex). You don't need them; Codex does.
</Note>