VECTORA INSTRUCTIONS

Your Complete Guide to Using the AI Detection Terminal

What is VECTORA?

VECTORA is a cutting-edge AI detection platform designed to verify the authenticity of digital content in the post-truth era. It's a neural firewall that helps you identify AI-generated text, images, and documents with high accuracy.

Our project combines three powerful AI providers (Google Gemini, Groq, and Cerebras) to give you the most comprehensive analysis possible. Whether you're checking a news article, verifying an image's authenticity, or analyzing a PDF document, VECTORA has you covered.

                🎯 Core Mission: In a world where seeing is no longer believing, verification is the
                new
                currency of trust. VECTORA empowers you to distinguish truth from synthetic reality.
            

Step-by-Step Guide to Using the Checker

Navigate to the Checker

Click on the "Checker" link in the navigation bar at the top of the page, or visit /check directly.

Choose Your AI Provider & Model

On the left sidebar, you'll see your current configuration. Click the "[ CHANGE CONFIG ]" button to open the settings modal.

Select Provider in Settings Modal

In the settings modal, use the "AI_PROVIDER" dropdown to choose between:
• Google Gemini - Best for all-purpose analysis
• Groq - Ultra-fast real-time processing
• Cerebras - Specialized for text analysis

Settings modal showing provider dropdown

Select AI Model

Click the "AI Model" dropdown to see available models. Each model displays capability icons:
📝 Text - Can analyze text content
📷 Image/Vision - Can analyze images
🌐 Docs/Web - Can search web and analyze documents

Apply Configuration

After selecting your preferred provider and model, click "APPLY CONFIG" to save your settings. The sidebar will update to show your active configuration.

Input Your Content

In the main console area, you have three options:
• Text Input: Type or paste text directly into the textarea
• File Upload: Click "[ UPLOAD DOC/IMG ]" to select an image or PDF
• URL: Paste a URL to an image or web page

Input area with textarea and upload button

Configure Web Search (Optional)

Toggle the "ENABLE WEB SEARCH" switch if you want the AI to perform live fact-checking against online sources. This is especially useful for verifying claims and detecting misinformation.

Note: If your selected model doesn't support web search, you'll receive a warning.

Execute Analysis

Click the "EXECUTE ANALYSIS" button to start the verification process. The system will:
• Validate your input
• Check model compatibility
• Process your content
• Stream results in real-time

Review Results

Results appear in an animated HUD (Heads-Up Display) below your input. You'll see:
• AI Probability Score: Percentage likelihood the content is AI-generated
• Detailed Analysis: In-depth reasoning and evidence
• Source Links: References to external verification sources (if web search enabled)
• SynthID Detection: Digital watermark analysis (Gemini models only)

Analysis results HUD with completed analysis

Clear Buffer (Optional)

To start a new analysis, click "[ CLEAR BUFFER ]" to reset all inputs and results.

Model Capabilities Legend

Understanding the icons you see in the model selection dropdown:

Icon	Capability	Description
TEXT	Text Analysis	Model can analyze and verify text content, including articles, paragraphs, and documents.
IMAGE	Vision/Image Analysis	Model can analyze images for AI-generation markers, deepfakes, and visual manipulation.
WEB	Web Search & Documents	Model can search the web for fact-checking and analyze PDF documents natively.

                💡 Pro Tip: Choose models with all three icons (📝📷🌐) for the most comprehensive
                analysis!
            

Tips & Best Practices

✅ For Best Results

• Provide sufficient context (at least 50-100 words for text analysis)
• Use high-quality images (avoid heavily compressed files)
• Enable web search for fact-checking news and claims
• Try multiple models if you get uncertain results
• Check the status bar for real-time processing updates

🔍 Understanding Results

• Scores above 70% indicate high likelihood of AI generation
• Scores below 30% suggest likely human-created content
• 30-70% range indicates uncertainty - consider context
• Always read the detailed reasoning, not just the score
• Cross-reference with external sources when possible

⚠️ Important Warnings

• AI detection is probabilistic, not definitive proof
• Always verify critical information independently
• Some models show compatibility warnings - heed them!
• External links show a warning modal for your safety
• Results can be inaccurate or hallucinatory - use judgment

Troubleshooting

Common Issues

Q: Why am I getting "MODEL INCOMPATIBLE" error?

A: You're trying to upload an image/PDF with a text-only model. Switch to a model with the 📷 icon (like Gemini 2.0 or Llama Vision).

Q: The analysis is taking too long. What should I do?

A: Try switching to Groq or Cerebras for faster processing. Large files and web search enabled will take longer with Gemini.

Q: My web search results seem irrelevant.

A: Web search works best with Gemini. Other providers may have limited or no web search capabilities. Check the warning message.

Q: What does "NO WEB INTELLIGENCE" warning mean?

A: The selected model cannot access live web data and is limited to its training data. Results may be outdated for recent events.

Q: Can I analyze multiple files at once?

A: Currently, VECTORA processes one file at a time. Clear the buffer between analyses using "[ CLEAR BUFFER ]".