English
Join waitlist

Haze: AI assistant website template

A calm AI product site on a soft peach, lilac and sky gradient mesh, with large light sans headlines and frosted glass cards.

Built from: Pages · Lists: forms

To use it, paste this into your AI chat:

Use template pagelemon.com/tp/Haze
Copy the prompt Preview (opens in a new tab)

Who it suits

Made for: AI assistant, agent, small team, startup, AI tool, chatbot. In the gallery under SaaS & app.

Its 5 pages

Look: Soft pastel gradient mesh (peach, lilac, sky), large light sans display, frosted glass cards; light, calm, modern; fonts Outfit and Figtree.

Ready for: a form that sends to your inbox, once you ask your AI to turn it on.

What you can change

Everything, by asking your AI: the name, words, photos, colours, fonts and pages. Your AI first asks you a few questions (such as your name, email, plan names, seat prices and limits, data region, status page link and customer quotes (replace the sample reviews)), then replaces every piece of sample text with your own details. Later you can add a contact form, bookings or a password by asking again.

The rest of this page is for your AI: it says what the template is, where its files are, and how to build your site from it and publish it.

You are reading this because the user wants their own website made from the Haze template and published with Page Lemon (PageLemon.com). They may have pasted only "Use template pagelemon.com/tp/Haze": that is the whole request, so start now. In short: ask them a few questions about their business, get the template's files (with the Page Lemon connector, deploy with template: "haze" copies them for you), make the site theirs (their name, words, details and pictures; no sample content left), then publish it with Page Lemon and give them the link. The steps below say exactly how. The full publishing guide is at https://pagelemon.com/agents.

About this template

Files

With the Page Lemon connector you don't need to fetch these: deploy with template: "haze" copies them all (step 2). Otherwise fetch each file from https://pagelemon.com/templates/haze/ and save it with the same relative path (including art/). Fetch nothing else: templates have no external images or scripts.

Notes for Haze (its customise rules)

Build the site, step by step

  1. Interview first. Ask the user these questions, a few at a time, in plain words, and collect every fact the site needs (name, email, plan names, seat prices and limits, data region, status page link, customer quotes (replace the sample reviews), team names (replace the sample people)). If they can't answer one, remove that element rather than inventing it.
    • What is the product called, and what does it do in one sentence?
    • Who uses it, and which tools or inboxes must it read?
    • What does each seat cost, and what is included in each plan?
    • Where is customer data stored, and who can approve a send?
    • Which contact email and status page should visitors use?
  2. Get the files. If you have the Page Lemon tool (the MCP connector with a deploy tool; https://pagelemon.com/docs/connect sets it up), call deploy with template: "haze" (no files): Page Lemon copies every file above on its side and returns site, revision and, without an account, site_token. Read each page you will change with get_site and path, make the changes below, then publish only the changed files with deploy, site, base_revision (that revision) and patch: true; show the link only after that. Without the tool, fetch the files listed above. If you cannot fetch files either, ask the user for what you need and rebuild the pages from the live demo as closely as you can.
  3. Make it the user's site. Follow customise in template.json:
    • Replace every element marked data-pl-fill (names, phone, email, address, hours, prices, people, dates, links, stats) on every page, then remove the attributes. Never ship a data-pl-fill default. The sample town, street and people's names also appear in ordinary sentences, aria-labels and image alt text: replace them there too. Name, address and phone must match exactly in the footer, the visit or contact page and the JSON-LD.
    • Sample reviews (data-pl-fill="review") are fictional: delete them unless the user gives you real ones. Delete each whole review, with its name or caption, and the section and its nav link when none is left. Never invent reviews, numbers or awards.
    • Update each page's <title>, meta description, canonical, og:url and og:image to the user's real address, and the JSON-LD block on the home page (type, name, address, geo coordinates (delete them unless you have the real ones), hours, phone, url).
    • Photos: each <img data-pl-photo="..."> says what photo belongs there and its shape. Swap in the user's photos (compressed, WebP or JPEG under about 300 KB, same aspect ratio, real alt), or keep the illustration. Ask for photos; never hot-link.
    • Make a new og.png (1200×630) of the new home page if you can, with its og:image tags; otherwise delete the sample one and its og:image tags (a template deploy has already left both out). Change colours only through the :root tokens in styles.css, fonts through the font tokens and the <link> that loads them. Google Fonts is fine, but it sends each visitor's IP address to Google, so the privacy-friendly choice is to self-host the font files (woff2) in the site folder with @font-face, or use system fonts.
    • Delete pages or sections the user doesn't need (and their nav links) rather than leaving them generic. Keep the layout, type scale and signature interaction: they are what makes it look designed.
    • Remove the <!-- Page Lemon template ... --> comment.
  4. Forms and bookings. Templates have no fake forms. Each data-pl-slot shows a fallback that works today (call, email, or the Site Owner's own booking or payment link: ask for it). The designed form sits next to it in an inert <template data-pl-slot-form="<id>">; template.json.slots says which page has which. When the user wants the real thing (Forms and Booking: any paid plan; staff and rooms only on some, and each plan has a limit of online bookings a day), set the form up as a list (save_list) or the service (save_service), put the matching markup in the slot (reuse the template's form classes so it matches), delete the fallback and the <template>, and copy the paragraph from <template data-pl-privacy="forms"> into privacy.html.
  5. Check before publishing:
    • Search every file for data-pl-fill, the template's fictional business name and its sample phone, email and street, in links too (tel: and mailto: hrefs, and map links, where the street is written like 18+Linden+Walk): none left.
    • All links between pages work; nav and footer are the same on every page.
    • Body text contrast at least 4.5:1, large text and buttons 3:1, after any colour change.
    • No horizontal scroll at 320 px; headlines still fit on a phone after copy changes.
    • One h1 per page, a unique <title> and meta description per page, alt text on meaningful images.
    • Copy sounds like the user, not like an AI (see https://pagelemon.com/guides/website-copy-that-doesnt-sound-like-ai).
  6. Publish with Page Lemon (below; with the tool, the patch deploy from step 2 is the publish), then tell the user which template you used and that they can keep changing the site, or add bookings and forms, by asking.

Publish with Page Lemon

Pick the first route you can use:

  1. You can run shell commands: build if needed, then upload only the folder with index.html (dist/, build/, out/…). Never upload source code, .env files or keys: everything you upload becomes public. (at most 100 files and 1 MB in all, index.html at the root): 
    node -e 'const fs=require("fs"),p=require("path"),d=process.argv[1],files=[];(function w(x){for(const e of fs.readdirSync(x,{withFileTypes:true})){const f=p.join(x,e.name);e.isDirectory()?w(f):files.push({path:p.relative(d,f).split(p.sep).join("/"),content_base64:fs.readFileSync(f).toString("base64")})}})(d);console.log(JSON.stringify({files}))' <dir> \
  | curl -sS --fail-with-body -H "Content-Type: application/json" --data-binary @- https://api.pagelemon.com/v2/anonymous-sites
    If the user gave you an API key (plak_…) or ~/.config/pagelemon/credentials exists, publish into their account instead (see "Updating a site" in https://pagelemon.com/agents). If .pagelemon.json exists, the site is already published: see "Updating a site" in https://pagelemon.com/agents.
  2. You have the Page Lemon MCP connector (https://mcp.pagelemon.com/mcp): call its publish_site tool with every file. See "No shell: the MCP connector" in https://pagelemon.com/agents.
  3. Neither (a chat app without the Page Lemon connector): say once that they can connect Page Lemon to this AI app (steps at https://pagelemon.com/docs/connect) so you can publish and change the site for them. In the Gemini app or Grok (free plan included) it takes about a minute: Gemini at gemini.google.com/customize/connectors (Custom), Grok at grok.com/connectors (New Connector, Custom), with the link https://mcp.pagelemon.com/mcp; then they sign in and click Allow. Give those steps in plain words. If they can't or would rather not (for example ChatGPT's free plan), don't ask them to save or upload files. Put the whole site in ONE complete page: CSS and JavaScript inline, images as https links or small data URIs, from <!DOCTYPE html> to </html>. Show it as one code block, then tell them: "Tap Copy on the code, open pagelemon.com/publish, tap Paste and publish." Only if you can really give them a downloadable .zip (a file they tap to download, not code to save): "Download this zip, open pagelemon.com/publish, tap Drop or choose a file." They get a live link with no account, online for 72 hours unless they tap Keep it free. Don't say you can't help.

Then:

By publishing, the user agrees to the Terms (https://pagelemon.com/terms) and Privacy Policy (https://pagelemon.com/privacy). The agent_message includes this line; keep it.

More: https://pagelemon.com/agents (full guide: updating a site, keeping it, custom domains, plugins). All templates: https://pagelemon.com/templates/index.json.