Hello, PageGun: The Marketing Site That Builds Itself

Most marketing websites are treated like projects.

You redesign them. You launch them. You promise everyone that this time the site will stay fresh. Then the product changes, the market shifts, the sales team needs a new page, the founder wants a launch post, support asks for better docs, and the changelog quietly becomes a graveyard.

That is not because teams are lazy. It is because websites were built around the wrong assumption: that humans should manually create and maintain every marketing surface.

PageGun starts from a different assumption.

Your marketing site should be alive.

It should research what pages need to exist. It should write the first version. It should build the page, attach the right metadata, route it correctly, and send it to a human for review. When the product changes, it should know which pages are now stale. When a campaign launches, it should produce the landing pages that campaign needs. When a feature ships, it should turn that update into proof that the product is moving.

That is what we are building.

PageGun is an AI marketing site builder

The familiar way to describe PageGun is "a website builder."

That is true, but incomplete.

Traditional website builders help humans assemble pages. PageGun helps teams operate a marketing site through AI agents. The point is not just to make page creation easier. The point is to make page creation part of the marketing system itself.

PageGun researches, writes, builds, and maintains pages for your go-to-market motion.

Those pages can be SEO pages, but they do not have to be. They can be landing pages for ads, comparison pages for sales, launch pages for new features, docs pages for developers, changelog entries for retention, outreach pages for a specific audience, or campaign pages for an email sequence.

The page is the unit of marketing work.

If a marketing idea cannot become a page quickly, it usually dies in Slack, Notion, or a half-finished doc. PageGun exists to close that gap.

The loop: research, build, approve, publish

The core loop is simple:

1. Research what should exist.
2. Draft the content.
3. Build the page.
4. Send it to a human for approval.
5. Publish only when it is ready.
6. Keep watching for updates.

The human approval step matters.

Autonomous publishing sounds impressive until it ships something dumb on your domain. PageGun is not trying to remove judgment from marketing. It is trying to remove the repetitive work around judgment: blank pages, formatting, routing, metadata, stale content checks, and the endless glue work between strategy and shipping.

The agent should do the boring eighty percent.

The human should make the call.

SEO is one motion, not the whole product

SEO is an obvious starting point because SEO is page-shaped.

A good programmatic SEO motion needs research, templates, structured data, internal links, page generation, reviews, publishing, and updates. That maps naturally to PageGun.

But reducing PageGun to an SEO tool would be too small.

The same system that can create search pages can create campaign pages. The same review flow that keeps SEO pages from becoming thin content can keep ads and outreach pages from drifting off-brand. The same content memory that helps maintain documentation can help a changelog explain why a feature actually matters.

Search is one traffic source. Ads, email, sales, partnerships, launches, and product updates are others.

PageGun is for the pages those motions need.

Changelogs are marketing too

A changelog is not just a list of commits wearing nicer clothes.

A good changelog proves the product is alive. It gives sales a reason to follow up. It gives customers confidence that the thing they bought is still improving. It gives search engines and AI systems fresh, specific product evidence. It gives your team a rhythm for turning shipping into storytelling.

That is marketing.

So PageGun should not only help create landing pages. It should help maintain the marketing surface around the whole product: docs, changelogs, launch posts, feature pages, comparison pages, and the connective tissue between them.

A product that ships every week should not have a website that looks like it froze six months ago.

What we are not building

PageGun is not a generic AI writer.

The world has enough blank text boxes with a sparkle button.

PageGun is also not trying to be a prettier CMS with an AI sidebar. A CMS stores content. PageGun should understand the job that content is doing: what campaign it supports, what audience it serves, what product truth it depends on, what pages it links to, and when it needs to be revisited.

And PageGun is not a replacement for taste.

Taste still matters. Positioning still matters. Strategy still matters. If anything, PageGun makes those things more important, because once the mechanical work gets cheaper, the quality of the judgment stands out more.

Why now

AI agents are getting good enough to do multi-step work.

They can research a market, inspect existing pages, draft content, generate page structure, call APIs, upload media, and prepare a publishable draft. But most websites were not built for agents. They were built for humans clicking through dashboards.

PageGun is built for the agent era.

It exposes the site as something software can work with: pages, articles, docs, media, metadata, routes, schemas, drafts, and publishing gates. That means an agent can do real work without pretending to be a person in a browser.

This is the boring infrastructure layer that makes the exciting agent workflow actually useful.

What this blog will cover

This blog will be where we think in public.

We will write about AI marketing sites, programmatic SEO, campaign pages, docs, changelogs, human-in-the-loop publishing, MCP workflows, Data Mode, and the product decisions behind PageGun itself.

Some posts will be strategic. Some will be technical. Some will be dogfood notes from using PageGun to build PageGun.

That is the point.

If PageGun works, this site should become its own proof: a marketing surface researched, drafted, built, reviewed, updated, and improved by the same system we are building for everyone else.

Hello, PageGun.

Now let's make the website move.