Calabi Labs · Guide · 2026-06-01
Fingerprintjs unveils ai assistant detection to filter chatgpt gemini
FingerprintJS AI Assistant Detection: Filter ChatGPT, Gemini & Claude Traffic in Real Time
FingerprintJS has launched a dedicated AI Assistant Detection capability that lets websites and applications identify, filter, and manage traffic coming from AI bots — including ChatGPT, Gemini, and Claude. The feature rolled out in June 2025 and is designed to give businesses full visibility into who (or what) is visiting their digital properties.
What It Does
AI assistants are increasingly used to scrape content, automate interactions, and run AI-powered fraud. Standard analytics can't tell you the difference between a real human visitor and a bot making an API call through ChatGPT. FingerprintJS's AI Assistant Detection closes that gap.
With this feature, you can:
- Detect AI bots visiting your site or app in real time
- Filter bot traffic from your analytics and reporting
- Block AI-powered scraping before it distorts your data
- Verify authorized agents with confidence scores
How It Works: IP-Range Matching + DNS Verification
FingerprintJS uses two complementary detection methods depending on the AI provider:
IP-Range Matching (ChatGPT & Claude)
FingerprintJS maintains verified IP ranges associated with OpenAI's ChatGPT and Anthropic's Claude. When traffic originates from one of these known IP blocks, the system flags it as an AI agent. This approach is fast, reliable, and works at the network level.
DNS-Based Verification (Gemini)
Google's Gemini behaves differently — it often resolves DNS before making a request. FingerprintJS detects this by checking DNS resolution patterns associated with Gemini's infrastructure. This method catches Gemini traffic that IP matching alone would miss.
For other AI assistants, FingerprintJS combines browser fingerprinting signals, network behavior patterns, and automation indicators into a single confidence score.
Why It Matters
| Problem | Impact |
|---|---|
| AI scrapers skewing analytics | Inaccurate visitor data, wasted ad spend |
| Content being pulled for AI training | Loss of proprietary content control |
| Bot-driven fraud and account creation | Revenue loss and security risk |
| SEO manipulation by AI-synthesized links | Damaged search rankings |
AI Assistant Detection addresses all of these by giving you the tooling to see and act on bot traffic — not just guess at it.
How Accurate Is It?
FingerprintJS claims high accuracy for recognized AI agents by cross-referencing multiple signals rather than relying on any single indicator. The system can distinguish authorized AI agents (like SearchGPT bots legitimately crawling your site) from unauthorized scraping bots. This is particularly valuable for publishers and content-heavy sites that need to comply with AI crawler's robots.txt directives while blocking bad actors.
How to Get Started
FingerprintJS AI Assistant Detection is available through the Fingerprint platform, including integration options via API and SDK. Organizations can:
- Enable AI agent detection in their Fingerprint dashboard
- Add a filter to exclude or tag AI traffic in analytics pipelines
- Set up real-time blocking or throttling for identified bot sources
- Combine it with other Fingerprint signals (browser fingerprinting, bot detection) for layered protection
Use Cases at a Glance
- Publishers — Keep AI crawlers out of your true user metrics
- E-commerce — Stop bots from scraping pricing and inventory data
- Financial services — Prevent AI-driven fraud and account creation
- SaaS platforms — Control which AI tools can access your API and content
What's Next
This launch is part of a broader FingerprintJS strategy called the Automation Intelligence API, which maps the full landscape of AI and automated traffic across the web. As AI assistants become standard browsers for LLMs, the need for robust detection tooling will only grow.
Try Calabi free at calabilabs.com — 3 cleans, no card.