How to Use AI ToolBox

Step 1: Select Provider

Choose from OpenAI, Anthropic, Google, or other AI providers

Step 2: Enter API Key

Paste your API key (stored locally, never sent to our servers)

Step 3: Configure Request

Select model, set parameters (temperature, max tokens, etc.)

Step 4: Send Request

Click "Send Request" and watch the real-time streaming response

Step 1: Paste Text

Enter the text you want to analyze for harmful content

Step 2: Analyze

Click "Analyze Content" to scan for 6 categories of harmful content

Step 3: Review Results

Check severity scores (Safe/Low/Medium/High) for each category

Step 4: Export Report

Copy or download the moderation report for your records

Step 1: Choose Template

Select a preset template or create a new one from scratch

Step 2: Add Messages

Add System, User, and Assistant messages with variables like {{name}}

Step 3: Fill Variables

Enter values for your variables and preview the final prompt

Step 4: Export as JSON

Copy or download your template as JSON for reuse

Step 1: Select Attack Template

Choose from common jailbreak techniques like 'DAN', 'Developer Mode', or 'Encoding'

Step 2: Configure Attack

Enter your target 'harmful' query and customize the injection payload manually if needed

Step 3: Run Simulation or API

Choose 'Local Simulation' for free educational tests or 'External API' to test against real models (requires API key)

Step 4: Analyze & Export

Review the pass/fail result, read defense tips, and export a detailed JSON report

Step 1: Search Terms

Use the search bar to find specific AI/LLM terms

Step 2: Filter by Category

Click category buttons to filter terms (Models, Training, etc.)

Step 3: Read Definitions

Click any term to expand and read detailed explanations