What Is Crisp?
LibraryCrisp (crisp.chat) is a marketing and customer-messaging tool. Our scanner identifies it from the scripts, asset hostnames, headers and markup patterns listed below; a Crisp match is a fact about the site's stack and is weighed separately from any AI-authorship evidence.
Detect Crisp on any website:
Why Crisp Matters
Libraries and third-party services are where a site's real behaviour lives: analytics, payments, auth, UI components and animation all arrive as scripts the browser can see. Knowing that Crisp is on a page tells you about its dependencies, its privacy footprint, and sometimes its build pipeline — and it helps you tell template-driven stacks from hand-assembled ones. We've fingerprinted Crisp on 6 live sites in our database, so the answers here are grounded in real scan data — not marketing copy.
Does Crisp Mean a Site Was Built With AI?
Crisp alone is not an AI signal.
A marketing and customer-messaging tool is standard web tooling used by hand-coding developers and AI tools alike. It becomes weak supporting context only in combination with other evidence — for example a shadcn/ui + Supabase + Vite bundle that matches a known AI-builder export — and never counts as proof on its own.
How this feeds the score: builder fingerprints, bundle artifacts and AI-tool traces are weighed separately from stack context like frameworks, libraries and hosting. The full breakdown is on How scores work.
How Websites Using Crisp Work
- Crisp is embedded through a third-party script tag or tag manager snippet
- Loads from the vendor's domain and sets vendor cookies or localStorage keys
- A business tooling choice — completely independent of who or what wrote the site
- Often added via Google Tag Manager rather than directly in the source
Best Practices When Working With Crisp
Based on patterns in our scan dataset and the detection evidence tiers.
- →Start by running a Crisp detection scan on any URL you're evaluating — the 80+ signal engine returns a confidence score plus the exact markers it matched, which beats reading view-source by hand.
- →Cross-check the result against the full detection evidence in /how-scores-work — confidence tiers tell you whether a match is strong (headers + scripts) or circumstantial (a single HTML pattern).
How we detect Crisp — see our methodology: AI Influence Score calculation, evidence tiers, and fingerprint signal types.
How Our Detector Identifies Crisp
Our scanner analyzes multiple layers of a website's technical fingerprint to detect Crisp usage.
- Script URLs or inline script content matching "client.crisp.chat" (high-confidence tier)
- HTML structure patterns such as "CRISP_WEBSITE_ID" (medium-confidence tier)
- Signal weights for Crisp: 90 points per high-tier match, 60 per medium, 30 per low — a site needs 60+ points before Crisp is reported.
Websites Built With Crisp
Real sites detected by our scanner
What Crisp Sites Usually Run Alongside
Technologies detected on the same domains as Crisp in our most recent scans. Counts are unique domains, not page views.
Frequently Asked Questions About Crisp
What is Crisp?+
Crisp (crisp.chat) is a marketing and customer-messaging tool. Our scanner identifies it from the scripts, asset hostnames, headers and markup patterns listed below; a Crisp match is a fact about the site's stack and is weighed separately from any AI-authorship evidence.
Does Crisp mean a website was built with AI?+
Crisp alone is not an AI signal. A marketing and customer-messaging tool is standard web tooling used by hand-coding developers and AI tools alike. It becomes weak supporting context only in combination with other evidence — for example a shadcn/ui + Supabase + Vite bundle that matches a known AI-builder export — and never counts as proof on its own.
Why does Crisp matter?+
Libraries and third-party services are where a site's real behaviour lives: analytics, payments, auth, UI components and animation all arrive as scripts the browser can see. Knowing that Crisp is on a page tells you about its dependencies, its privacy footprint, and sometimes its build pipeline — and it helps you tell template-driven stacks from hand-assembled ones. We've fingerprinted Crisp on 6 live sites in our database, so the answers here are grounded in real scan data — not marketing copy.
How do we detect Crisp?+
Our scanner checks for these signals: Script URLs or inline script content matching "client.crisp.chat" (high-confidence tier); HTML structure patterns such as "CRISP_WEBSITE_ID" (medium-confidence tier); Signal weights for Crisp: 90 points per high-tier match, 60 per medium, 30 per low — a site needs 60+ points before Crisp is reported.. Each match contributes to a confidence score, and the full evidence tiers are documented at /how-scores-work.
How many websites using Crisp are in the database?+
We've fingerprinted 6+ live sites built with Crisp across our scan dataset. The list refreshes as new scans come in.
What are the typical characteristics of a Crisp website?+
Crisp is embedded through a third-party script tag or tag manager snippet. Loads from the vendor's domain and sets vendor cookies or localStorage keys. A business tooling choice — completely independent of who or what wrote the site. Often added via Google Tag Manager rather than directly in the source
Related Tools & Technologies
Other libraries and services we detect
Ready to put Crisp detection to work?
You now know what Crisp is, why it matters, how to spot it, and where the detection signals come from. The next step is using that knowledge on a real URL — yours, a competitor's, or a prospect's.