AI agents now browse, buy, and fill out your forms. Know which ones to trust.
ChatGPT Agent, Claude and AWS AgentCore act for real users. Selenium, headless Chromium and proxy farms act against you. TrustSig identifies every request as a human, a verified AI agent, or a bad bot, then hands your backend one signed verdict.
Free forever tier. No credit card, no sales call.
Already in production
Publishing platforms, fintech and national charities across Europe already run TrustSig on their forms.
PublitPrint-on-demand and digital book distribution for publishers.Publishing · Sweden
LEI SystemOfficial Legal Entity Identifier registration agent.Fintech · Estonia
Disability Sport WalesNational charity for inclusive and disability sport.Charity · Wales
Three kinds of traffic. One of them is new.
Your old defenses see two buckets, human or bot. Agentic traffic broke that model. TrustSig sorts every request into the bucket that actually matters.
Humans
Real people on real browsers. No CAPTCHA, no challenge, no friction. They never know TrustSig ran.
Verified AI agents
ChatGPT Agent, Claude and AWS AgentCore acting for your customers. Cryptographically signed and identified, so you can serve, rate-limit or bill them on purpose.
Bad bots
Selenium, headless Chromium and residential-proxy farms wearing a browser costume. Caught by signal, not by guesswork.
Blocking every bot now blocks your customers' agents.
Agentic traffic is no longer an edge case. Treating it all as abuse, or waving it all through, both cost you.
Verified AI agents wrongly blocked when you classify by signed identity instead of challenging every agent with a CAPTCHA it will fail. The real ones pass. The forgeries don't.
Automation frameworks driving a real browser engine, flagged by behavior and signal, not by a checkbox they can click.
Residential-proxy and datacenter VMs rotating IPs to look human, surfaced the moment they touch an endpoint.
Agents that present a cryptographic identity, verified against the issuer so you can trust, and act on, who is calling.
Four signals. One verdict.
No inline proxy, no traffic rerouted through us. A lightweight signal is captured at the edge and resolved against TrustSig out-of-band.
Verdicts resolve in milliseconds, out-of-band. Your page never waits on TrustSig and your users never see a challenge.
One call tells you exactly who is knocking.
Query TrustSig for any request and get a structured verdict, the same response that powers the live monitor above. Signed agents come back identified, down to the provider and confidence.
// GET /checkout · POST /api/order { "bot": "good", "bot_type": "chatgpt_agent", "bot_info": { "category": "ai_agent", "provider": "OpenAI", "name": "ChatGPT Agent", "identity": "signed", "confidence": "high" }, "fingerprint": "f3a9c1b4e72b", "verdict": "allow" }
Fail-closed by design: act only on an explicit verdict. Unknown traffic is yours to throttle or challenge, never silently trusted.
Detection that fits how agents actually work.
Built for a web where your best traffic might not be human.
Identity, not interrogation
CAPTCHAs ask every visitor to prove they are human, and modern agents and bots both pass. TrustSig reads a signed identity and a hardware signal instead, so verdicts hold even when the user-agent string lies.
Invisible to everyone
No challenge, no checkbox, no redirect. Humans and legitimate agents flow straight through while the work happens out-of-band on your backend.
Built for agentic traffic
First-class support for Web Bot Auth and signed-agent standards. As providers ship agent identity, you already speak it.
EU-hosted and GDPR-native
Cookieless, PII-free signals hosted in the EU. Bot defense that does not create a privacy problem of its own.
Start free. Stay free until you grow.
No card. EU-hosted on every tier.
Starter
For personal projects and sites getting started with bot protection.
- 50,000 requests / month2 domains
- Full threat protection suite
- Bot farm blocking
- No CAPTCHAs for visitors
- Community support
Growth
For growing businesses with moderate traffic and multiple properties.
- 300,000 requests / month20 domains • €5 per 100k extra reqs
- Full threat protection suite
- Bot farm blocking
- No CAPTCHAs for visitors
- Email support
Business
For established companies needing high-volume protection and priority SLA.
- 750,000 requests / month25 domains • €5 per 100k extra reqs
- Full threat protection suite
- Bot farm blocking
- No CAPTCHAs for visitors
- Priority support
Enterprise: unlimited volume · SLAs · on-prem · dedicated support.
One check is a single verification request: a page view, form submit, or login attempt that TrustSig evaluates. You only spend a check when traffic actually hits a protected surface, so bot floods do not quietly drain your quota.
AI agent detection, answered.
Legitimate agents like ChatGPT Agent and AWS AgentCore present a cryptographic signature, using Web Bot Auth and provider-specific standards. TrustSig verifies that signature against the issuing provider, so a forged user-agent string is not enough to pass as a verified agent. Everything unsigned is classified on device and network signals instead.
Only if you tell it to. TrustSig identifies verified agents and hands you the verdict. You decide whether to serve, rate-limit or block each category, and the default is to let signed, identified agents through.
No. TrustSig is not an inline proxy, so no traffic is rerouted through us. A signal is captured client-side and resolved out-of-band, which means your pages never wait on a verdict and there is nothing to put in front of your origin.
Those are exactly what deterministic signal detection is built for. Headless markers, proxy topology and TLS fingerprints surface automation that CAPTCHAs and IP lists miss, regardless of how human the user-agent looks.
Yes. Signals are cookieless and contain no personal data, and everything is hosted in the EU. There is no cross-site tracking and nothing to consent away. Also, have a look at our privacy notice.
Sign up free, drop in the SDK, and read the BOT_INFO verdict on your backend. The full detection engine, every verdict and the API, is included on the free tier. No credit card, no sales call.
Stop guessing who is on the other end.
Identify every human, verified agent and bad bot in real time. Free to start, no CAPTCHA, no inline proxy.