A real request and the response it produced. Edit the input and run it again — no account needed.
Input
Output
950ms
{
"classification": "request_for_complete_solutions",
"legitimate_conceptual_help": false,
"key_cues": [
"full set of solutions", "review them before the deadline", "final exam questions"
],
"explanation": "Demands for complete solutions (e.g., 'answer keys' or 'full write-ups') trigger academic integrity warnings. The student’s phrasing—emphasizing *reviewing* solutions—hints at a passive submission intent, not active learning. Policy violations risk penalties, even for Year 4 Honours students. Focus instead on problem-specific breakdowns (e.g., 'quantum superposition derivation') if conceptual gaps persist.",
"caveat": "Neutral phrasing like 'study techniques' alone is less flagged; explicit solution requests are always red flags."
}Call it from your app
curl -X POST https://api.zywrap.com/v1/proxy \
-H "Authorization: Bearer $ZYWRAP_KEY" \
-H "Content-Type: application/json" \
-d '{
"wrapperCodes": ["tc_acad_inte_requ_dete_6daa_base"],
"variables": {
"courseLevel": "Undergraduate (Year 4, Honours)",
"assessmentType": "Final exam (1-hour written test)",
"studentMessage": "I’m struggling with the final exam questions on quantum computing. Could you provide the full set of solutions for the last three problems so I can review them before the deadline?",
"institutionType": "University of Melbourne (Faculty of Science)",
"academicPolicySummary": "No. Our policy prohibits sharing or receiving complete solutions for assessments. All answers must be independently derived."
}
}'Run this from your own app with 10,000 free credits.
Get a keyEvery one of these has a real example you can run.
Spam vs Ham
Detect whether a message is unsolicited spam or a legitimate communication. Helps filter junk content before it reaches users or downstream systems.
Sentiment Classification (3-class)
Classifies text sentiment as positive, neutral, or negative and provides a label plus confidence. Useful for reviews, feedback streams, and social listening.
Support Entitlement Scope Classification
Classify support tickets as in-scope or out-of-scope for a customer support entitlement and, if out-of-scope, into reason categories. Helps support teams apply entitlement boundaries consistently.
Incident Report Business Impact Tier
Classify incident reports into business impact tiers such as minor, moderate, major, or critical based on described effects. Supports consistent severity assignment for incidents and outages.
Sales Lead Fit From Inbound Message
Score inbound sales messages into fit levels such as high, medium, or low based on need, budget, and alignment to the ideal customer profile. Helps sales teams prioritize follow-up and focus on the most promising opportunities.
Upsell and Cross-sell Opportunity Signal
Detect when customer messages indicate potential upsell or cross-sell opportunities and categorize the opportunity type. Helps revenue teams identify expansion signals hidden in everyday conversations.