Yes — Codex can build, launch and manage your cold email campaigns in WarmySender, all in plain language, through the same 78-tool connection Claude, ChatGPT, Cursor, Codex, OpenClaw, and Hermes Agent share. You ask in the Codex terminal. It never sends a message itself and can never raise a limit: it writes the campaign and hands it to WarmySender’s scheduler, which paces every email, LinkedIn, Instagram and WhatsApp action inside safe limits no matter who is driving. Connecting and disconnecting accounts stays in the app.
Codex is OpenAI’s coding agent and runs in your terminal, so you talk to it in the shell you already use for the rest of your pipeline. It connects with one add command and one login command; the login opens a WarmySender sign-in in your browser, so there is no key to paste. Cold email through Codex suits the operator who scripts the rest of the pipeline — build the list, ask Codex to create the campaign from it, start it, and read the report in the same shell. Setup is one add command and one login.
Every "can" below stands on a real tool in the WarmySender connection. Every "cannot" is not a setting — there is simply nothing an agent can call that does it.
| Can / cannot | What that means for cold email |
|---|---|
| Can | Create a multi-step campaign, edit it, and set the sending window and daily limits. |
| Can | Start, pause and resume a campaign — starting only hands it to the scheduler. |
| Can | Add prospects one at a time or in bulk, update their details, and enroll or unenroll them. |
| Can | Work from your saved email templates instead of writing every email from scratch. |
| Can | Read and add to your suppression list so the wrong people are never contacted. |
| Can | Report on every campaign and explain why nothing went out today. |
| Can | Add a branded tracking domain and check that it is set up correctly. |
| Can | Set the postal address and opt-out wording used in your email footer. |
| Cannot | Send an email itself — it writes the campaign and WarmySender’s scheduler sends at your safe pace. |
| Cannot | Raise a daily cap, a warmup stage or a domain ceiling — nothing it can call moves a limit. |
| Cannot | Connect or disconnect a mailbox — credentials never pass through an agent. |
| Cannot | Switch off the opt-out in your footer — it can change the style, never remove it. |
| Cannot | Read or reply to responses — replies stay in WarmySender’s inbox, with you. |
OpenAI's coding agent registers WarmySender with one command and signs in through your browser — no key to paste.
Run both lines in your terminal; the second opens your browser to sign in to WarmySender.
codex mcp add warmysender --url https://warmysender.com/mcp
codex mcp login warmysenderPrefer a key? Create one under Settings › API Keys and choose exactly which permissions it carries, including read-only. With a key, the Codex route is: Run both lines, adding the export to your shell profile so it sticks.
Both routes reach the same workspace and stay inside the same safe limits. Every agent's steps live on the AI Agents hub.
Create a three-step campaign called "Agency founders" for this list, weekdays only, from my two warmed mailboxes, then start it.
Pause every campaign that uses the sales@ mailbox until Monday.
Why did the "Q4 partners" campaign send nothing yesterday?
Add these 40 people to the suppression list and take them out of any running campaign.
Read prospects.csv, create a three-step campaign from it, weekdays only, and print the campaign id once it has started.
Codex never sends a message itself and can never raise a limit. Creating or launching a campaign only writes it and hands it to WarmySender's scheduler, which paces every email, LinkedIn, Instagram and WhatsApp action inside safe limits and a gradual ramp — the same protections that apply when a human clicks the buttons. Connecting and disconnecting accounts stays in the app. Account safety always wins.
Yes. Connect Codex to WarmySender and ask in plain language, in the Codex terminal — it can build, launch and manage your cold email campaigns. It never sends a message itself and can never raise a limit: it writes the campaign and hands it to WarmySender’s scheduler, which paces every email, LinkedIn, Instagram and WhatsApp action inside safe limits no matter who is driving. Connecting and disconnecting accounts stays in the app.
OpenAI's coding agent registers WarmySender with one command and signs in through your browser — no key to paste. Run both lines in your terminal; the second opens your browser to sign in to WarmySender. The command text is: codex mcp add warmysender --url https://warmysender.com/mcp codex mcp login warmysender — the server address is https://warmysender.com/mcp. The steps are on this page under "Connect Codex", and the full install hub for every agent is at /ai-agents.
No. Nothing an agent can call moves a limit — not a mailbox’s daily cap, your domain’s ceiling or the pace across the sending window. Codex can read where you are against a limit and tell you when a pause lifts; WarmySender’s scheduler decides what actually goes out.
No. Connecting and disconnecting accounts stays in the app on purpose — a credential should never pass through an agent or a chat log. Connect the mailbox once in WarmySender; after that Codex can use it.
Nothing goes out in a burst. Starting a campaign only queues it, and WarmySender’s scheduler drains that queue at your normal daily pace across your sending window, inside each mailbox’s limit and your domain’s ceiling. Pausing is one sentence and is never blocked, and everything the agent built is a normal campaign you can open, edit or delete.
No. The add command registers the server address https://warmysender.com/mcp and the login command opens a browser sign-in — nothing to paste or store. Prefer a key anyway? Create one under Settings › API Keys and use the key form on the AI Agents hub.
No. Codex only writes the campaign or the setting; once it has handed that to WarmySender, the scheduler paces every email on its own whether or not Codex is running.
Yes. Each client registers the server and signs in on its own; they reach the same workspace and the same safe limits, and the scheduler paces every email exactly the same whichever agent wrote the campaign.
Last updated: 2026-09-09