Connect WarmySender to Muse Code

Set it up in three steps

1. Create a WarmySender key. In WarmySender, open Settings → API Keys and create one (it starts with ws_). Grant everything for full control, or only the read permissions if you want your agent to look but not touch. The key is shown once — copy it somewhere safe.
2. Add WarmySender to Muse Code's settings file. Open ~/.config/muse/settings.json — create the folder and the file if they are not there yet — and put this in it, replacing the placeholder with your own key:

   {
     "schema_version": 1,
     "mcp_servers": {
       "warmysender": {
         "transport": "streamable_http",
         "url": "https://warmysender.com/mcp",
         "headers": {
           "Authorization": "Bearer ws_YOUR_API_KEY_HERE"
         }
       }
     }
   }

   Save this as ~/.config/muse/settings.json (create the folder if it is not there yet) and start Muse Code again. Keep the schema_version line — Muse Code rejects the file without it. If you already have a settings file, add only the warmysender block inside your existing mcp_servers.

3. Start Muse Code again and ask for something. "List my campaigns and tell me which ones sent nothing yesterday." If that answers, you are live.

If your computer uses a custom configuration folder, Muse Code looks for the same file inside that folder instead — the file name and the contents do not change.

Why the first line matters

Muse Code will not read a settings file that does not declare which format it is written in. The schema_version line at the top is that declaration, and it has to say 1. Leave it out, or change the number, and Muse Code refuses the entire file — not just the WarmySender part — so anything else you had configured stops working too. If Muse Code starts up complaining about your settings file, that line is the first thing to check.

The second thing to check: a WarmySender entry describes a server it reaches over the network, so it carries an address and headers and must not also carry a command to run. If you copied another server's block as a starting point and left a command line in it, remove that line.

Why a key, and not a sign-in screen?

Several agents let you connect by signing in to WarmySender in your browser, with no key to look after. Muse Code does not offer that for MCP servers today — a key in the settings file is the route it supports, so it is the only route we publish for it. Nothing is lost: a key reaches exactly the same workspace, stays inside exactly the same safe sending limits, and can be revoked at any moment in Settings → API Keys. A key is also the better fit for a terminal agent in the first place, because there is no browser window waiting for you to click approve.

Using an agent that does offer the sign-in route? Connecting an AI agent that asks you to sign in walks that path end to end.

Treat the key like a password: it lives in a file on your own machine, so do not paste it into a chat, a screenshot or a shared repository. If it ever gets out, delete it in Settings → API Keys and create a new one — nothing else needs changing.

What you can ask Muse Code to do

• Build and launch cold email campaigns — a multi-step sequence, a sending window, daily limits, then start it.
• Build and launch LinkedIn, Instagram and WhatsApp campaigns — and pause, resume and enroll people on request.
• Find and add prospects — search the built-in business database, save what it finds into a list, import a batch, enroll them.
• Verify email addresses — one address or a whole list, before you commit to sending.
• Tune warmup — turn it on for a mailbox, set its strategy and target volume, and watch its reputation build.
• Report — how a campaign did, who replied, which mailboxes look weak, and why something sent nothing yesterday.

The safety line that never moves

Your agent never sends a message, DM or invite itself, and it can never raise a limit. Creating or launching a campaign only writes it down and hands it to WarmySender's scheduler, which paces every email and every LinkedIn, Instagram and WhatsApp action inside safe daily caps and the gradual ramp — the very same protections the app uses, whether a human or an agent pressed go. Ask it to enroll a huge list at once and the result is a queue that drains at your normal safe pace, not a burst. Email verification shares the same allowance and pacing. And connecting or disconnecting your accounts stays in the app — your mailbox and social logins never pass through an agent. Account safety always comes first.

Frequently asked questions

Where is Muse Code's settings file?

At ~/.config/muse/settings.json. Create the folder and the file if they are not there yet. If your computer is set up with a custom configuration folder, Muse Code looks for a muse/settings.json inside that folder instead — same file name, same contents. When Muse Code complains about your settings on startup, the message names the exact file it read, which is the quickest way to confirm you edited the right one.

Muse Code says my settings file is not valid. What is wrong?

Nine times out of ten it is the schema_version line. Muse Code requires it, it has to say 1, and without it the whole file is refused — including anything else you had configured. Copy the block on this page exactly, keeping that first line. The other common cause is a leftover command line inside the WarmySender entry: WarmySender is reached over the network, so the entry carries an address and headers and must not carry a command.

Can I sign in with my WarmySender account instead of using a key?

Not in Muse Code today — it has no browser sign-in step for MCP servers, so a key in its settings file is the route it supports. That is the whole difference: a key reaches the same workspace, obeys the same safe sending limits, and can be revoked whenever you like in Settings → API Keys. Agents that do offer a sign-in screen are covered in our guide to connecting an agent that asks you to sign in.

I already have other servers in my settings file. Will this replace them?

Only if you overwrite the file. Add the warmysender entry inside your existing server list rather than pasting the whole block over the top, and keep your existing schema_version line as it is. Everything you already had keeps working.

I added it but no WarmySender tools appear. What now?

Start with the settings file itself: Muse Code refuses the whole file if the schema version line is missing or wrong, so nothing will load. Next, check the key — if it was mistyped, or if it was deleted in WarmySender, the connection is turned down at startup. Next, check your plan: connecting an agent needs an active paid plan and is included at no extra charge, and on a free account the connection attaches but every request it makes is refused. Finally, if the key you created carries read-only permissions, reading works and everything else stays hidden — create a key with the permissions you need. Restart Muse Code after any change so it reloads the tool list.

Can Muse Code send my cold emails?

No agent sends anything, and that is deliberate. Muse Code builds the campaign, enrolls the people and starts it — and starting it simply hands it to WarmySender's scheduler, which paces every send inside safe daily caps and a gradual ramp. It also cannot raise a limit, and it cannot connect or disconnect a mailbox or a social account — that always stays with you in the app.

Does this cost extra?

No. Connecting an AI agent is included with any active WarmySender plan — there is no separate add-on to buy.

Related reading

• Connect your AI agent — the agent-agnostic start-here guide, with the steps for Claude, ChatGPT, Cursor, Codex, OpenClaw, and Hermes Agent
• Connecting an AI agent that asks you to sign in — the sign-in route, for the agents that offer it
• Which AI agents can run outreach — the per-app, per-plan comparison table
• What your AI agent can and cannot do with cold email — the capability matrix
• Connect an AI agent — one-paste setup for every supported agent, and where to create your key
• The WarmySender MCP server — the full tool reference