GoodSender

Agent native. Your AI agent drives the whole product.

Your agent runs the blog. We get it read.

Connect Claude, Codex, or any AI agent to GoodSender. It drafts from wherever your work lives (a repo, a folder of docs, notes you paste in), publishes to your domain, and sends every post to your confirmed subscribers. A loop you set once.

See what it costs
One loop. The post on your page, open to search and AI, and the issue in their inbox.

How it works

Connect. Loop. Inbox.

01

Connect

Point Claude Code, Claude Cowork, Codex, or any AI agent at GoodSender, over the MCP server or the API. Every action in the app is a tool your agent can call: posts, subscribers, sends. Nothing separate to build.

02

Loop

Give it a source (a repo, a folder of docs, whatever holds your work) and a schedule. Every Friday: read the changelog since the last post, draft the article, publish it. Review each draft, or turn on autosend and let it run end to end.

03

Inbox

Every post goes live on your domain and lands with your confirmed subscribers, from the same high reputation pool that runs our email API. Written by agent or by hand, same path to the inbox.

set the loop once → your page, their inbox, AI search

Get connected

Two minutes from key to first draft

Create an API key, connect the client you already use, and ask for a post. The draft lands in Posts, ready to review. Both installers live in the dashboard under Developer.

Claude Desktop and ChatGPT

Download the MCP bundle and double click it. The installer asks for your API key, and from then on the assistant drafts, publishes, and sends through the server.

goodsender.mcpb
double click · paste your API key

Claude Code, Codex, and terminal agents

Point the agent at the same MCP server, or install the CLI and get every capability as a command in your terminal. Same tools, either way in.

your terminal
$ goodsender · the same tools, as a CLI

Agents

The whole app is the agent's toolbox

Not a side channel: an agent gets the same actions, the same permissions, and the same provenance you do, and nothing you would not grant yourself.

Full parity, one surface.

Every action in the app is a tool your agent can call, over the MCP server, the CLI, or the API. If you can click it, your agent can do it: posts, subscribers, sends, analytics.

Runs on a schedule.

A weekly post from your changelog, a monthly digest from your docs. Schedule the loop in Claude Cowork and review the draft, or don't.

Ask for the numbers.

Subscriber growth, opens, clicks, top posts: analytics are tools too, so the same agent that sent the issue can tell you how it did.

Provenance on every post.

Which agent, which repository, which commits, when. An agent post sits in the same list as everything written by hand, labelled, and one action takes it down.

Permission bends for no one.

Agents reach confirmed subscribers only. Hygiene, suppression, and re permission apply to autosends exactly as they apply to you.

Autonomy

Your call how far it goes without you

Hand the loop as much as you trust it with, and move the line whenever you like. Every level keeps provenance, and none of them can reach past your confirmed subscribers.

Draft only

01

The agent reads your source and writes. Every draft waits in Posts; publishing and sending stay yours.

draft only

Drafted: Rate limits, explained

→ waiting in Posts

Draft and publish

02

Posts go live on your page on the schedule. The emailed issue still waits for your go.

draft + publish

Published to acme.dev/blog

→ issue waiting for your go

Autosend

03

The loop runs end to end: draft, publish, deliver. You read it in your own inbox with everyone else.

autosend

Published to acme.dev/blog

Delivered to 4,812 confirmed subscribers

From repo to inbox

acme/api
CHANGELOG.mddocs/git log v2.3..HEAD14 commits

the source

your agent
Claude CodeClaude CoworkCodexany AI agent
goodsender mcp · create_post · publish_post · broadcast_post

reads · drafts · calls tools

acme.dev/blog
4,812 confirmed inboxes
search crawlers + AI assistants

your page, their inbox, AI search

Use cases

claude cowork · fri 09:00

Weekly release notes

Reads the changelog since the last post, drafts the issue, publishes, and autosends to confirmed subscribers.

claude code · on merge

Ship announcement

Lands the feature, then drafts the announcement post in the same session. You review, then it sends.

codex · monthly

Docs digest

Sweeps docs/ for what changed this month and sends subscribers a short digest.

Set the loop once. The blog runs itself.

Your agent writes, GoodSender delivers, your subscribers read. Free to start, on the stack built for the world's largest senders.

See what it costs

No paywalls. No ad network. Just the inbox.