Passive real-time interception
Guardrail watches live calls and messages the instant they arrive — no prompt to write, no conversation to re-tell. Protection that works even when you don’t suspect anything.
Real-time scam interception
Stay safe,
spot scams instantly.
Guardrail analyses your live calls and messages as they happen, surfaces the red flags a scammer is counting on you to miss, and tells you exactly how to respond — while you’re still on the line.
On-device analysisNothing recordedCancel any time
LiveOn-device
Unknown caller · +1 (555) 0192
Claiming to be “Fraud Department”
“We’ve detected suspicious activity on your account.”
“To secure it, read me the 6-digit code we just texted you.”
“Stay on the line — don’t hang up or the transfer completes.”
Asks for a one-time passcode
Manufactured urgency
Discourages hanging up
Bank impersonation scamrisk 96
Hang up now. No real bank asks for a one-time code. Call the number printed on your card to verify.
Illustrative product demo
3,500
Closed-beta subscribers
1.2s
Median time to first red flag
12
Languages at launch
0
Conversations stored on our servers
Sample figures from our closed beta
The gap
Scams don’t fail because you’re careless. They work because they’re fast.
A convincing voice, a countdown, a link that looks almost right. By the time you’d think to open an app and describe what happened, the money has already moved.
Guardrail removes that step entirely. The check happens automatically, in the seconds that actually matter.
Bank & fraud-desk impersonation
Fake delivery and customs fees
Tech-support remote access
Investment & romance payouts
How it works
Set up once. Protected from then on.
01
Turn it on once
Grant Guardrail the call and message permissions during a 40-second setup. That is the last time you have to think about it.
02
It listens in the background
Every incoming call and message is analysed on-device as it happens. No pasting, no screenshots, no describing what happened.
03
You get a verdict, not a lecture
When something looks wrong you get a risk score, the exact red flags detected, and one clear sentence telling you what to do next.
Features
Built for the moment the scam is happening.
Not a report you read afterwards. Not a chatbot you have to remember to open.
Guidance in the moment
A flag alone doesn’t help under pressure. Guardrail scripts your next move: what to say, what to refuse, and which official number to call back on.
Pattern detection, not blocklists
Blocklists only catch numbers that already burned someone. Guardrail reads tactics — passcode requests, urgency, payment redirection — so brand-new scams are caught on day one.
On-device by default
Analysis happens in a rolling buffer on your phone that is discarded when the conversation ends. Nothing is recorded, nothing is uploaded unless you report it yourself.
Protect the people you worry about
Add a parent or relative to your plan and receive an alert summary when they’re targeted — a heads-up, never a transcript of their private calls.
One price, every market
$6 a month, worldwide, covering up to five devices. Safety utilities shouldn’t cost more in the places where scams do the most damage.
Positioning
Passive interception, not another chat window.
Today’s scam tools split into two camps: chatbots that need you to describe the scam, and blocklists that only know numbers other people already reported. Guardrail is neither.
| Capability | Guardrail | AI scam chatbots | Caller-ID blocklists |
|---|---|---|---|
| Works without you noticing anything is wrong | Yes | No | Yes |
| Reads the content of a live conversation | Yes | No | No |
| No pasting, forwarding or describing required | Yes | No | Yes |
| Catches unknown numbers on their first use | Yes | Yes | No |
| Tells you how to respond, not just whether it’s bad | Yes | Yes | No |
| Covers calls and messages in one place | Yes | No | No |
Category comparison. AI scam chatbots include tools such as Scamio and Norton Genie; caller-ID blocklists include Hiya and Truecaller. Categories describe how each approach works, not a feature audit of any one product.
Early testers
What the first 3,500 people told us.
“The alert came up while the phone was still ringing. I’d have handed over the code — it sounded exactly like my bank’s fraud line.”
“I installed it for my dad. Three weeks later it flagged a “tech support” call and told him to hang up. He did. That was the whole point.”
“I’d tried a scam-checker chatbot before. I never remembered to open it. Guardrail just sits there and does the checking for me.”
Placeholder testimonials — fictional, for layout purposes
Consumer SignalThe Daily LedgerFraud WeeklyNorthwind ReviewSafe & Sound PodcastBytewise DigestConsumer SignalThe Daily LedgerFraud WeeklyNorthwind ReviewSafe & Sound PodcastBytewise Digest
Press placeholders — fictional publications
Pricing
One price. Everywhere.
A safety utility should cost about the same as a coffee, whether you’re in Oslo or Nairobi.
Watchlist
Message scanning only.
Free
- Text & link scanning
- Basic risk score
- 1 device
- Community scam feed
Most popular
Guardrail
Full real-time interception.
$6/ month
- Live call + message interception
- Red flags with response guidance
- Up to 5 devices
- On-device analysis, nothing stored
- Weekly pattern updates
Family
For households and relatives.
$12/ month
- Everything in Guardrail
- Up to 5 protected people
- Trusted-contact alert summaries
- Shared trusted-number list
Billing opens at launch. Joining the waitlist costs nothing and commits you to nothing.
Be protected before the next call.
Join the waitlist and we’ll send you an invite as Guardrail rolls out region by region.
FAQ
Questions people ask first.
No. Analysis runs on your device in a rolling buffer that is discarded the moment a conversation ends. Nothing is written to disk, and nothing is uploaded unless you explicitly tap “Report this scam”.
Chat tools only help after you notice something is wrong and go to the trouble of pasting or describing it. Guardrail runs passively in the background, so the warning arrives while the call is still ringing — which is exactly when pressure tactics work.
Patterns rather than blocklists: requests for one-time passcodes, manufactured urgency, lookalike domains, payment redirection, remote-access requests, and impersonation of banks, couriers, tax offices and support desks. New patterns ship weekly.
Rarely. Guardrail stays silent unless a conversation crosses a confidence threshold, and you can tune sensitivity per channel. Contacts you mark as trusted are never flagged.
Guardrail is launching on iOS and Android with a desktop browser extension to follow. Interception ships in 12 languages at launch, with more added as regional scam patterns are mapped.
Safety shouldn’t be priced like enterprise software. One global price keeps Guardrail affordable in every market we serve, and a single subscription covers the whole household — up to five devices.
Yes. Family plans let a trusted contact receive an alert summary when a high-risk interaction is flagged — without ever seeing the conversation itself.