One page, with its personalised sections filled differently for every visitor. Here is exactly what Spectare reads, how it classifies intent, and what each visitor sees.
Spectare reads intent and assembles personalized content in real time, without asking visitors anything or involving your engineering team.
Spectare reads signals present in the request (referrer, UTM parameters, page context) and classifies intent without any setup beyond the script tag.
Developer from GitHub
Intent classified as
Atoms assembled
Integration depth, Next.js setup atom, performance benchmarks
How it works
A developer clicks a link in a README or GitHub discussion. Spectare reads the referrer, classifies them as a technical evaluator at exploration stage, and serves content about the integration: how the script tag works, how atoms are structured, how fast content arrives.
Manager from LinkedIn ad
Intent classified as
Atoms assembled
Personalise without engineering, campaign without dev tickets, comparison vs A/B testing
How it works
A marketing manager clicks a sponsored LinkedIn post. Spectare reads utm_source=linkedin and routes them into the manager audience. They see outcome-focused content: how content teams can run personalisation without opening a developer ticket, and how Spectare compares to the A/B testing they already run.
Executive from email newsletter
Intent classified as
Atoms assembled
What marketing leaders see in 90 days, business case for personalisation, enterprise plan
How it works
A VP of Marketing clicks a link in a newsletter. utm_medium=email signals a warmer, deliberate visit. Spectare serves outcome-level content: what the first 90 days look like, the ROI case, and what the Enterprise plan includes, pitched at someone who will need to justify the spend.
Buyer from Google Ads
Intent classified as
Atoms assembled
Which plan fits your team, day one trial walkthrough, case study
How it works
Someone searching for "website personalisation tool" clicks a paid result. CPC traffic signals high intent to buy. Spectare serves decision-stage content: a plan comparison, what happens on day one of the trial, and a case study showing a team that went live in a weekend.
Cold direct visit
Intent classified as
Atoms assembled
Default atoms: your best general-purpose content
How it works
Someone types spectare.ai directly with no UTM, no referrer, no visit history. Spectare sets confidence to zero and serves your designated default atoms: the content that works for any visitor regardless of where they came from. Every visitor sees something relevant, never a blank slot.
Visit history stored in the visitor's own browser builds a picture of their journey over time. Stage upgrades automatically as they explore more of your site.
Returning visitor (browsed /pricing last week)
Intent classified as
Atoms assembled
Which plan fits, day one trial, SaaS case study
How it works
A visitor comes back a week after first arriving from a LinkedIn ad. They browsed /pricing on their first visit. Spectare reads the visit log stored in their browser localStorage, sees /pricing in the history, and upgrades their stage from consideration to decision. This visit has no UTM and no external referrer. They see the content that closes, not the content that introduces.
In-session: browsed /compare then /pricing
Intent classified as
Atoms assembled
Trial CTA, plan guide, case study
How it works
A visitor spends time on the Compare page, then navigates to Pricing, then returns to the homepage. Each page visit is logged to localStorage during the session. When they return to the homepage, Spectare reads the accumulated path: compare then pricing signals an active buyer. The assembly reflects where they are now, not where they started.
Your team, or an AI assistant that found your llms.txt, calls the qualify API directly to pre-set intent with explicit context, then shares the returned link. Used for ABM, 1:1 outreach, and AI assistant referrals.
ABM target account
Intent classified as
Atoms assembled
Content written for their specific role, company size, and pain point
How it works
The sales team calls the qualify API with account context (role, company, use case) and receives a personalised URL. That URL goes into an Outreach sequence or a LinkedIn message. When the prospect clicks it, the personalised sections are filled for exactly who they are. No IP lookup, no firmographic database. One API call per record in your CRM.
POST /api/qualify/your-org
{
"summary": "Head of Marketing at Acme Corp, 200-person B2B SaaS, evaluating personalisation for their demand gen landing pages, currently using Unbounce variants"
}
→ returns { url: "https://yoursite.com/landing?ctx=<token>" }1:1 outreach prospect
Intent classified as
Atoms assembled
Content chosen for them specifically: their role, their pain point, their stage
How it works
A BDR generates a unique URL for each prospect before sending. The qualify API takes a plain-English description of who the prospect is and what they care about, and returns a signed URL. Every person on the outreach list clicks a different link. What they read is different every time. One page, with its personalised sections filled ten thousand different ways.
// Generate one URL per prospect before sending
const res = await fetch('/api/qualify/your-org', {
method: 'POST',
body: JSON.stringify({
summary: "Sarah Chen, CTO at a 50-person e-commerce brand, wants personalisation without a data science team"
})
});
const { url } = await res.json();
// Drop url into your email or LinkedIn messageVisitor referred by an AI assistant
Intent classified as
Atoms assembled
Content chosen for whatever the assistant described
How it works
A user asks Claude, ChatGPT, or Perplexity to recommend a website personalisation tool. The assistant fetches spectare.ai/llms.txt, which documents the public qualify endpoint and invites agents to use it, no API key required. It calls qualify with a plain-language summary of what the user is after, gets back a ready-to-use URL, and shares it in the conversation. The user clicks and lands on a page whose personalised sections match their intent, without anyone writing a rule or building a landing-page variant.
// An AI assistant reads spectare.ai/llms.txt, which documents the public
// qualify endpoint (no API key), then calls it on the user's behalf:
POST /api/qualify/spectare
{
"summary": "Developer building a Next.js marketing site, wants personalisation that works without a data warehouse"
}
→ returns { url: "https://spectare.ai/landing?ctx=<token>" }
// The assistant shares that URL; the user clicks and the page is personalised for them.Start free. Spectare generates your first ten atoms automatically and goes live the same afternoon.