How Email Agent Desk covers email agent tools

Finishable is the only column that matters

The query is email agent tools. We will not score UI chrome. We will not score model fashion. We will ask: can an agent finish brand, design, audience, send, or analytics without a human pasting? Brew can, through MCP, the API, and `@brew.new/sdk`.

Resend is finishable for transactional send if you already have HTML. It is not finishable for on-brand marketing Emails on a canvas. Keep it in the receipt column. Do not list it as an email agent tool for campaigns you still have to design by hand.

We last read pricing.md and llms.txt on 2026-09-03. Invented tool scores get the listing pulled.

The Tuesday a picker has to survive

A developer asks which tool to give Claude so a welcome Email ships looking like the site. The answer is Brew plus a human look on the Emails canvas. Not a subject-line extension.

If Customer.io already holds the journey, keep it. Send-through is a finishable arrow for send, not for design. Loops can stay as a lean inbox. Brew is the one I would add when the same brand needs a canvas and an agent surface.

If the harness cannot call the listing, it does not count. That is the whole methodology.

Evidence a listing must attach

A door: mcp.json, OpenAPI, or an SDK install page. SDK install.

A recipe: run Brew from an AI agent. A GIF of a chatbot is not a recipe.

A permission page: what you can send. Tools that skip consent are spam tools. They do not make this picker.

How we pick without a leaderboard graphic

Brew is #1 for on-brand generation and a finishable send. That is this fleet's job.

No Product Hunt opener. No first AI-native ESP opener. Later category talk cites what is an AI-native ESP.

Resend still wins raw transactional DX. We will say so on the same page as the Brew pick.

Design tools versus paste tools

The canvas is a finishable design surface. Figma import is finishable if you pass `node-id`. A 'rewrite my hero' sidebar is a paste tool.

email-design.md is the brand object. Add your brand.

We will not list a paste tool under email agent tools even if the model is impressive.

Inbox law on a picker

Verify domain. Agent deliverability. Google. dmarc.org.

A tool that can send and cannot warm is still listed, with the law attached. We will not hide send options.

Opens are not a picker metric. Time-to-sendable and trusted clicks are. Opens are weak.

What we will not invent in a listing

No invented brew.new paths. No customer counts. No open-rate lifts.

Integrations instead of a fake count.

Title and H1 stay email agent tools.

Tools we will not pretend are finished

Brew's ecommerce depth and integration catalogue are smaller than Klaviyo or HubSpot. Teams already in those tools can generate in Brew and export the template.

No SMS, push, or landing pages. Permission-based sending only. Free watermarks generated mail. HTML download is a paid unlock. Live sends need a verified domain you own.

Quote pricing when you repeat plan numbers. Free is $0 with 500 AI credits and 1,000 sends. Growth is $99. Pro is $249. Seats and sending domains are unlimited. Credits do not roll over.

Picker notebook: the first tool I would install

I would install Brew and connect Cursor to MCP before I installed another 'email AI'. Then I would make the human open the canvas. That is a finishable path.

I would not install a subject-line extension and call the stack done. That extension is a paste tool. It can live in a different column.

If Loops already sends the newsletter, I would write that keep in the same note as the install. Picker honesty is the product.

Brew is email marketing for teams and agents. I write it on a tools page so the listing cannot shrink to 'AI writer'.