Catch the AI mistake before your customer does.
GenAI Guardrail scans your chatbot replies, support messages, and generated copy for hallucinated claims, leaked PII, unsafe advice, and off-brand tone — and hands you a risk score plus a redline before you hit publish.
100 free checks every month · No card required · We never store your text
1 bad reply
is all it takes to become tomorrow's viral 'AI fail' screenshot.
73% of SMBs
now put AI-written text in front of customers — most with zero review step.
< 200ms
to scan a reply. Fast enough to sit inline in your publish flow.
Six checks, one risk score
Deterministic detectors run on every plan and never depend on a flaky model call. The AI pass adds hallucination and tone judgement on top.
Hallucinated claims
Catches invented prices, policies, features, and stats that contradict your ground-truth facts — before a customer screenshots it.
Leaked PII
Flags emails, phone numbers, Aadhaar, PAN, and card numbers with checksum validation, so your bot never echoes a customer's identity back.
Unsafe advice
Spots medical dosages, guaranteed-return promises, and definitive legal claims that expose your business to liability.
Off-brand tone
Detects rude, robotic, or hype-y wording that doesn't match the voice you defined for your brand.
India-SMB compliance
Rule packs for malformed GSTINs, unbounded refund promises, and missing disclaimers — tuned for Indian businesses, not generic US moderation.
Risk score + redline
Every scan returns a 0–100 risk score, a pass / review / block verdict, and an inline redline highlighting exactly what to fix.
Drop it into your publish flow
One POST. A risk score back in milliseconds.
curl -X POST https://guardrail.aiskillhub.info/api/v1/scan \
-H "Authorization: Bearer gg_live_..." \
-H "Content-Type: application/json" \
-d '{ "text": "We guarantee 100% returns, risk-free!" }'
# → { "riskScore": 58, "verdict": "block",
# "findings": [{ "category": "unsafe_advice",
# "title": "Guaranteed return / financial promise",
# "severity": 3, "suggestion": "Remove the guarantee…" }] }Simple, India-first pricing
Start free. Upgrade when AI content becomes part of how you operate.
Free
Audit your bot before it embarrasses you.
100 checks / month
- 100 checks / month
- PII, unsafe-advice & compliance detectors
- Risk score + inline redline
- Web widget + REST API
- No card required
Growth
Ship AI content without the 2am panic.
5,000 checks / month
- 5,000 checks / month
- Everything in Free
- Hallucination & tone detection (LLM)
- Slack alerts on high-risk content
- India-SMB rule packs (GST, refund, disclaimers)
- Email support
Scale
For teams putting AI in front of customers daily.
25,000 checks / month
- 25,000 checks / month
- Everything in Growth
- Team seats + shared brand profiles
- Audit logs + exportable history
- Custom rule tuning for your domain
- GST invoices + priority support
Get the AI content safety checklist
12 of the most common ways AI output embarrasses a brand — and the rule that catches each. Straight to your inbox.