OpenAI Codex
Add staqly to the Codex CLI so it can write posts and put them on the calendar.
Connect the Codex CLI to staqly as an MCP server, and Codex can write posts, schedule and publish them, and answer when a schedule asks for more, just from your prompts. You set it up once per machine.
Connect to Staqly Community Edition
Start staqly
Run staqly serve (or the Docker image). See the Quickstart.
Add the MCP server
Your agent sends staqly's access token, which staqly prints at every start. Put it in
STAQLY_API_KEY (in your shell profile, so the agent always has it):
export STAQLY_API_KEY=$(staqly serve token) # Docker: $(docker exec staqly server token)Codex reads it from that variable on every request:
codex mcp add staqly --url http://localhost:3333/mcp --bearer-token-env-var STAQLY_API_KEYOr merge this into ~/.codex/config.toml:
[mcp_servers.staqly]
url = "http://localhost:3333/mcp"
bearer_token_env_var = "STAQLY_API_KEY"Tip: staqly mcp config codex prints the command and the TOML for your current
CLI profile.
STDIO (Docker only)
With Docker you can use STDIO instead. It runs mcp-stdio inside the staqly container,
so it needs no token, and it does not work with staqly serve:
codex mcp add staqly -- docker exec -i staqly mcp-stdioConnect to Staqly Cloud
Codex reads the bearer token from an environment variable:
export STAQLY_API_KEY=stk_...
codex mcp add staqly --url https://app.staqly.ai/mcp --bearer-token-env-var STAQLY_API_KEY[mcp_servers.staqly]
url = "https://app.staqly.ai/mcp"
bearer_token_env_var = "STAQLY_API_KEY"Then ask: "Draft a post about the release into Product updates and schedule it for tomorrow 9:00 on my X channel." More in Example prompts.