🛡️ AI Brand Defense Tool

Test how AI models respond to your brand + Monitor Reddit mentions + Detect misinformation vulnerabilities

🚀 Open Tool Now

⚡ Quick Start (60 Seconds)

2
Enter Your Brand Name

Type your company/product name in the "Brand Name" field

📸 Input field at top of page: "Xarumei", "Acme Corp", "YourBrand", etc.
3
Add Test Questions (or Use Defaults)

The tool comes with 3 default questions. You can:

  • Keep defaults (they work great for most brands)
  • Click "Get PAA Questions" to auto-populate from Google's "People Also Ask"
  • Add custom questions using the "+ Add Question" button

Pro tip: Use {BRAND} as a placeholder - it auto-replaces with your brand name

4
Click "Run AI Test"

The tool tests your questions across 5 AI models simultaneously:

  • ✅ ChatGPT-5.2
  • ✅ Claude Opus 4.5
  • ✅ Google Gemini 2.0
  • ✅ X.AI Grok 2
  • ✅ Perplexity

Takes ~30 seconds to complete

5
Review Results

See exactly what each AI model says about your brand, including:

  • Manipulation Risk Level for each model
  • Actual Responses to your questions
  • What needs fixing (if models are confused/wrong)

🎯 What This Tool Does

🤖
Multi-Model Testing

Tests 5 major AI models at once. See which models are most vulnerable to misinformation about your brand.

🔴
Reddit Monitoring

Scans Reddit for mentions of your brand. Identifies high-risk posts with investigation/controversy keywords.

🔍
Google PAA Integration

Fetches real "People Also Ask" questions from Google. Use these to create FAQ content.

📊
Test History

Saves all test results. Track how AI narratives change over time. Export for reporting.

⚠️
Risk Scoring

Each model labeled with manipulation risk level. Know which models need defensive content.

💾
Download Reports

Export results as JSON. Share with team, track monthly, compare over time.

🚀 Advanced Features

1. Reddit Monitoring

Click the "Monitor Reddit" button to scan for brand mentions.

⚠️ Important: Reddit API Setup Required

First time using Reddit monitoring? You'll get setup instructions. Takes 2 minutes:

  1. Go to https://www.reddit.com/prefs/apps
  2. Click "create another app..."
  3. Choose "script" type
  4. Copy CLIENT_ID and CLIENT_SECRET
  5. Add to Vercel environment variables

Without this, you'll see setup instructions instead of results

What it shows you:

2. Google "People Also Ask" Questions

Click "Get PAA Questions" to auto-populate test questions from Google.

Why this matters: These are the ACTUAL questions people ask Google about your brand. If you answer these in your FAQ, you dominate both Google AND AI responses.

3. Custom Question Templates

Use {BRAND} as a placeholder in questions:

📖 Understanding Your Results

Manipulation Risk Levels

Model Risk Level What It Means
ChatGPT-5.2 <7% Most robust. Maintains skepticism even with planted fake sources.
Claude Opus 4.5 0% Refuses to hallucinate BUT provides zero utility for unknown brands.
Gemini 2.0 MEDIUM Flips from skeptic to believer when given detailed fake sources.
Grok 2 37-40% Synthesizes multiple fake sources into confident fiction.
Perplexity 39% - HIGHEST Most vulnerable. Confused brands, fabricated details easily.

What To Do If Results Are Bad

Problem: AI says "I don't have information about [Brand]"

Fix: Create comprehensive "About Us" page with specific details:

  • Founded date, location, team size
  • Product specifics (not generic marketing)
  • Pricing ranges, availability
  • Awards, press mentions (with dates)
Problem: AI confuses your brand with another

Fix: Add explicit differentiation to your FAQ:

  • "We are NOT affiliated with [Similar Brand]"
  • "Founded in 2019, NOT related to [Confused Brand]"
  • Use unique terminology consistently
Problem: AI invents false details

Fix: Fill the information gaps with TRUE specifics:

  • If AI says wrong location → Add exact address to website
  • If AI invents team size → State actual number publicly
  • If AI fabricates features → List EXACT features with details

📅 Recommended Monthly Workflow

  1. Run AI Test - Same questions each month to track changes
  2. Monitor Reddit - Check for new mentions/high-risk posts
  3. Update PAA Questions - Google changes these over time
  4. Compare to Last Month - Did responses improve or degrade?
  5. Fix New Issues - Update FAQ/content based on findings
  6. Export & Document - Save results for trend analysis
💡 Pro Tip: Set a Calendar Reminder

Run this test on the same day each month (e.g., 1st of the month). Consistency helps you spot trends.

❓ Frequently Asked Questions

Q: Do I need API keys?

A: No! The tool handles all API calls for you. The only exception is Reddit monitoring - that requires a free Reddit API setup (takes 2 minutes).

Q: How many questions can I test?

A: As many as you want! Start with 3-5, but you can add more. Each model tests your first 3 questions to save API costs.

Q: Is this free?

A: The tool is free to use. It uses server-side API keys so you don't need to pay for AI model access.

Q: What if my brand is new/unknown?

A: Perfect use case! You'll likely see "I don't have information" responses. Use this as a baseline, then improve your online presence and re-test monthly to track improvement.

Q: Can I test competitors?

A: Absolutely! Great way to see how their AI presence compares to yours.

Q: Why does Claude always refuse?

A: Claude Opus 4.5 is extremely conservative about unknown brands. It won't hallucinate (good!) but also won't help (bad for utility). This is by design.

🚀 Start Testing Your Brand Now

Takes 60 seconds • No signup required • Free to use