How to Use
Aiscern
Step-by-step guides for every detection modality. Learn how to get the most accurate results and interpret what the scores mean.
How to Detect AI-Generated Text
A step-by-step guide to identifying AI-written content using Aiscern's text detection ensemble. Covers paste, upload, URL scan, and interpreting confidence scores.
- 1Paste or upload your text
- 2Review overall confidence score
- 3Examine sentence-level heatmap
- 4Export report if needed
How to Verify Deepfake Images
Learn how to upload images for AI generation detection, understand the pixel-integrity analysis, and interpret results for synthetic images from Midjourney, DALL-E, Stable Diffusion, and more.
- 1Upload image (JPG, PNG, WebP)
- 2Review ViT classifier confidence
- 3Check pixel integrity signals
- 4Review per-generator fingerprint analysis
How to Detect AI-Cloned Audio
Understand how Aiscern analyzes audio files for voice synthesis signatures. Guide covers supported formats, spectral analysis, and interpreting voice clone detection results.
- 1Upload audio file (MP3, WAV, M4A)
- 2Wait for spectral analysis (8–15s)
- 3Review voice synthesis confidence
- 4Check ASVspoof benchmark comparison
How to Detect Deepfake Video
A complete guide to video deepfake detection using Aiscern's frame-level ViT ensemble and temporal consistency analysis. Covers face-swap, lip-sync, and full-body synthesis detection.
- 1Upload video file (MP4, MOV, WebM)
- 2Frame extraction and analysis begins
- 3Temporal consistency check runs
- 4Review frame-level confidence breakdown
How to Use Batch Processing
Speed up bulk content verification with Aiscern's batch analysis. This guide covers how to upload multiple files, interpret batch results, and export bulk reports.
- 1Navigate to /batch (Pro+ required)
- 2Drag and drop multiple files
- 3Monitor job progress dashboard
- 4Export results CSV or individual PDFs
Understanding Confidence Scores
A deep dive into what Aiscern's confidence percentages actually mean, how to interpret uncertain results, and when not to act on a single detection result.
- 1≥62% = AI classification
- 239–61% = Uncertain zone
- 3≤38% = Human classification
- 4Always use alongside human judgment
Need more help?
Check our FAQ, read the API docs, or contact our support team.