AI can make the pages. Deciding what they’re for is still your job.

Aesthetary teaches the judgment around AI web production. You research the business, decide what the site has to do, plan its pages, choose a visual direction, build with an agent and check the result before anyone sees it. Each step ends in a file you keep, so you and your AI tool work from the same decisions.

Fast is the easy part

An agent will build whatever it’s asked for, quickly and with confidence. What it’s asked for is usually vague. A brief says editorial and means four different things to four readers; asked for Swiss Style, a model reproduces a grid and a sans-serif and misses the logic underneath, because nobody wrote that logic down.

The same gap runs through the whole project: who the site is for, what it has to prove, how its pages are organized, which claims are true. When those decisions live in someone’s head, the agent guesses them again in every session. The site looks done on launch day. A month later nobody can extend it, check it or explain it.

What you learn, and what you leave with

Each stage has a page and its own skills. You leave each one with files in your project that the next stage, and the next agent session, can read.

  • PrepareCLAUDE.md or AGENTS.md
  • Researchresearch/claims.md, research/words.md
  • Planbrief.md, requirements.md
  • Structureia/site-plan.md, ia/content-model.md
  • Designdirection.md, tokens.json
  • Buildthe code, with its Git history
  • Testqa/record.md, qa/gate.md
  • Improvemaintain/routine.md
The Research stage page: what's already in the repository, the four files Research writes, and what Strategy writes next.
The Research stage: what it needs, what it writes, what it hands on.

Who it’s for

Anyone making websites with AI doing some of the work. Designers and developers who want the craft to survive the speed. Freelancers and agencies who need work they can explain to a client. Business owners briefing and checking their own site. And the people who teach it, from universities to workforce programs.

Find the route for your role

How it fits together

The research feeds the design library, the library feeds the skills, and the skills sit inside the workflow you follow on a real site.

  1. Research

    Visual languages read source by source. Every claim says whether it’s documented or our own reading, and gaps stay visible.

  2. Design intelligence

    28 Aesthetics ready to build, the Elements they’re made of.

  3. Learn

    Skills for each stage, on the judgment the tools can’t make for you.

  4. The workflow

    The stages themselves, each ending in files a person and an agent can both read.

Annie Vickrey, founder of Aesthetary
Annie Vickrey, founder

Who started it

Aesthetary was founded by Annie Vickrey, a digital producer, designer and builder whose work spans digital strategy, UX, operations, websites and product development.

She came to it through making digital things: years of moving between agencies, digital production, operations, UX and, more and more, hands-on web production. AI changed how much of that chain one person can cross, and it made the decisions between the steps matter more. Aesthetary is where she writes those decisions down.

Authority should come from showing the work. Not performing certainty.

Annie Vickrey

Speaking at AllWeb, Tirana, November 6, 2026 (Annie’s speaker page, opens in a new tab)

Where it’s heading

Today everything on the site is free: the stages, the skills, 28 Aesthetics ready to build, Resources and Saved. Only Projects needs an account. A free one keeps your ticks, worksheets and projects on every device you sign in on.

Talent shows the websites people have built, for those who agree to be listed. The research keeps going one language at a time, and corrections from people who know a tradition better than we do are the most useful thing anyone can send.

What works so far

Three ways in