Landing-page work should start with intent
Page Owners should be able to describe an offer, audience, and change request without assembling a canvas from scratch.
About LP1
Let agents run the landing-page loop without losing the boundary.
LP1 exists for Page Owners who want agents to build, test, and operate Page Sets without giving up control of Page HTML, evidence, access, or Publish.
Why LP1 exists
Landing pages should be delegated without disappearing into a builder.
The product is built around one simple belief: agents can do the heavy work, but the Page Owner should still understand the evidence, constrain the action, and control Publish.
Agents need real source to work well
LP1 keeps Page HTML as the source of truth so External AI Clients can inspect and edit the Page precisely.
Publishing should stay part of the loop
Preview, Publish, conversion data, Experiments, Form Submissions, and webhooks stay close to the Page instead of becoming separate setup projects.
Product principles
A calm surface for Page Owners, an agent-ready platform underneath.
Surface
Chat is the friendly front door
The Page Owner path starts with natural language. Page HTML and Preview stay available, but they support the creation loop.
Source
Page HTML remains yours
LP1 avoids a proprietary page model. The Page can be inspected by people and operated by agents without losing the document underneath.
Backbone
MCP gives agents a stable way to operate
External AI Clients can create Pages, inspect Page Sets, find outliers, prepare Experiments, and run owner-approved automation through capability-scoped access.
Trust model
The Page stays inspectable from first draft to Published URL.
LP1 does not ask Page Owners to trust a hidden page format. The system keeps the Page, agent access, and capture workflow legible.
Start buildingPreview before Publish
Draft work can be reviewed through Preview before it becomes a Published URL.
Evidence stays attached to the Page Version
LP1 records views, Form impressions, conversions, Experiment Arms, and delivery receipts in the same operating surface.
External AI Client access is explicit
Owner-controlled External Client Connections make agent access visible instead of hidden in a black box.