# Connect Codex to Tesselith

> Connect OpenAI Codex to Tesselith and schedule social posts from the Codex CLI or the Codex IDE extension.

Canonical: https://tesselith.com/ai-agents/codex

**Before you start:** You need the Codex CLI or IDE extension and a Tesselith workspace on Pro, Team or Agency (every new workspace starts with a 14-day Pro trial) with at least one channel connected. The CLI, the IDE extension and the ChatGPT desktop app share one config file, ~/.codex/config.toml, so you configure Tesselith once.

## Quick setup

```bash
# 1. Add the Tesselith MCP server
codex mcp add tesselith --url https://api.tesselith.com/mcp

# 2. Sign in through your browser
codex mcp login tesselith

# 3. Check it
codex mcp list
```

## Step by step

### 1. Add the server

codex mcp add writes an entry to ~/.codex/config.toml. You can also add it by hand.

```toml
[mcp_servers.tesselith]
url = "https://api.tesselith.com/mcp"
```

### 2. Sign in

Run codex mcp login tesselith. A Tesselith window opens. Sign in, pick the workspace, untick anything you don't want the agent to do, and click Allow access. On a remote machine, add --no-browser to get a link you can open elsewhere.

### 3. Verify inside Codex

Start codex and run /mcp to see Tesselith and its tools, then ask Codex to list your channels. Tools load when a session starts, so restart after changing the config.

### 4. Use it from the IDE extension

The Codex IDE extension reads the same config.toml. Restart your editor after adding Tesselith and it's available there too.

## With an API key

For a server or CI job, point Codex at an environment variable holding your key. In Tesselith, open Settings → Developers, create an API key and tick only the scopes the agent needs. Keys start with tsl_live_ and are shown once, so store it in your password manager or a secret store.

```toml
[mcp_servers.tesselith]
url = "https://api.tesselith.com/mcp"
bearer_token_env_var = "TESSELITH_API_KEY"

# then, in the shell that runs codex:
# export TESSELITH_API_KEY=tsl_live_…
```

## Troubleshooting

- **Startup times out:** Raise startup_timeout_sec = 30 under [mcp_servers.tesselith].
- **401 with an API key:** TESSELITH_API_KEY isn't set in the shell that launched Codex, so the header is empty. Export it and restart.
- **Browser sign-in can't reach Codex:** Use codex mcp login tesselith --no-browser on remote machines.
- **Tools don't show up:** MCP tools load at session start. Restart Codex after editing config.toml.
- **An older Codex rejects url =:** Update Codex. Very old versions needed experimental_use_rmcp_client = true for HTTP servers.

## Questions people ask

### Does Codex cloud use my Tesselith connection?

Codex cloud doesn't read your local config.toml. Use the CLI or IDE extension, or connect Tesselith to ChatGPT.

### Where does Codex store MCP settings?

In ~/.codex/config.toml, shared by the CLI, the IDE extension and the ChatGPT desktop app. Trusted projects can also have .codex/config.toml.

### Should I use OAuth or an API key?

Use the browser sign-in on your own machine. Use an API key with limited scopes on servers and in CI.

Docs: https://learn.chatgpt.com/docs/extend/mcp?surface=cli

---
Tesselith · https://tesselith.com · Full site index: https://tesselith.com/llms.txt