# Get LinkedIn sales with Codex
> Use OpenAI Codex as a coding agent on Omentir. Put MCP in config.toml. Keep the token in an env var. Draft in the session. Send from the workspace. Codex is not ChatGPT chat.
- HTML: https://omentir.com/use-cases/codex-outbound
- Markdown: https://omentir.com/use-cases/codex-outbound.md

**Verdict:** Codex is useful when you already run OpenAI's coding agent against the repo. It is not the ChatGPT connector. Put LinkedIn in Omentir. Put the token name in bearer_token_env_var.

Codex reads TOML. Omentir sends. You still take the meeting.

## Highlights

- config.toml, not a connector UI
- Token stays in an env var
- Review list, then send

## First weeks

1. **Finish Omentir in the browser.** LinkedIn connected. My Product written. Codex will not fix a slogan.
2. **Add the server in config.toml.** Global file at ~/.codex/config.toml, or a project file only if the repo is trusted. url is https://omentir.com/api/agent/v1/mcp. bearer_token_env_var is OMENTIR_API_KEY, the variable name, not the token.
3. **Confirm with /mcp, then a scored list.** If the server is missing, fix the table name and the env var. Fetch agents.md. get_context, then list_agents. Pull up to 30 people with drafts. Do not send.
4. **Send from Omentir.** Cut junk. Start a small campaign. Close the session and the work stops.

## Why the TOML file matters

The CLI and the IDE extension share the same file. Get the table name wrong (mcp-servers instead of mcp_servers) and Codex ignores the block. ChatGPT chat still uses Settings and Connectors.

A project-local .codex/config.toml only loads for trusted projects. If tools never appear, start with the global file.

## Paste this into Codex

Export the env var before you launch. Fetch agents.md. Replace the brackets. Keep the last two sentences.

```
You are Codex, running against my repo through Omentir.

My product helps [buyer] get [result]. Best-fit accounts are [company type, size, region]. Strong signals: [signal]. Skip [bad fit].

Work only through the Omentir MCP server named in config.toml. Read https://omentir.com/agents.md if you have not. Start with get_context, then list_agents. If no finder matches this ICP, show me a create_agent config and wait for a yes.

Do not put secrets in config.toml. bearer_token_env_var is the name of an environment variable, not the token.

Pull up to 30 people. For each one write fit 1-5, the evidence, any risk, and a two-sentence draft that cites a real trigger. If a note could fit two buyers, rewrite it.

Do not send. Do not enroll. Do not sign into LinkedIn.
Do not create, delete, or reply unless I ask in this session.
```

## Related

- [Codex integration](https://omentir.com/integrations/codex.md): config.toml, /mcp, and the ChatGPT split.
- [Codex for LinkedIn outreach](https://omentir.com/blogs/codex-linkedin-outreach.md): Longer walkthrough of the TOML job.
- [How do I connect Codex to Omentir?](https://omentir.com/help/how-do-i-connect-codex-to-omentir.md): Short help version of setup.

## Frequently asked questions

**Is Codex the same as ChatGPT?**

No. ChatGPT uses workspace approval. Codex uses TOML and a Bearer env var.

**Can I put the API key in the TOML file?**

You can. You should not. The file is easy to commit.

**Should I pick Cursor instead?**

Pick the agent you already open. Cursor has its own MCP client. Codex shares config.toml across CLI and extension.

## Connect Codex only if that agent is already in the repo

If you wanted a chat tab, use the ChatGPT connector.

[Create an Omentir account](https://omentir.com/signup). Pro is $49/month.