How can we help you,
program your environment?
Get Started
Writing Prompt
Edge Hardware
Intergration
- Mastering Syntax
- Defining Spatial Polygon
- Setting Cooldowns
- Handling false positives
- Mastering Syntax
- Defining Spatial Polygon
- Setting Cooldowns
- Handling false positives
- Mastering Syntax
- Defining Spatial Polygon
- Setting Cooldowns
- Handling false positives
- Mastering Syntax
- Defining Spatial Polygon
- Setting Cooldowns
- Handling false positives
- Help Center
- Writing NLP Rules
- Mastering Syntax
Mastering NLP Rule Syntax
Argu Inc. (“Argu”, “we”, “us”, or “our”) respects your privacy and is committed to protecting it through our compliance with this Privacy Policy. This policy describes the types of information we may collect from you or that you may provide when you visit the website argu.ai (our “Website”) and use our edge-based computer vision platform, APIs, and associated services (collectively, the “Services”), and our practices for collecting, using, maintaining, protecting, and disclosing that information.
Please read this policy carefully to understand our policies and practices regarding your information and how we will treat it. By accessing or using our Services, you agree to this Privacy Policy.
The Anatomy of a Perfect Prompt
While Argu can understand conversational English, structuring your commands clearly will yield the highest precision and the lowest false-positive rates. We recommend structuring your rules using the following formula:
Examples of Excellent Prompts
Alert me if an unauthorized vehicle enters the loading bay and lingers for more than 5 minutes.Detect any person walking through the server room corridor without a visible ID badge.Track forklifts exceeding speed limits in warehouse sector 4.
Pro Tip: Specificity is Key
Instead of saying “Watch for bad guys,” define exactly what constitutes an anomaly. “Alert if a person accesses the roof door after 8:00 PM” provides exact temporal and spatial constraints that the vision model can execute flawlessly.
Handling Confidence Thresholds
By default, the Argu Edge Node will only trigger an alert if it is 85% confident that your conditions have been met. If you are monitoring a high-security zone (like a bank vault), you may want to lower this threshold to ensure you catch every possible event, even if it increases false positives.
You can append confidence overrides directly in your natural language prompt:
“Detect any person near the perimeter fence. Set confidence threshold to 60%.”
Reviewing Compiled Output
Once you submit a rule in the dashboard, you can click on the “View Compilation Log” button to see exactly how the LLM translated your request. It will show you the exact object classes (e.g., class_id: 0 (Person)) and spatial polygons it has locked onto.
If the compilation looks incorrect, simply rephrase your prompt and deploy again. Rule updates take less than 2 seconds to propagate to your Edge Nodes.