LLMWhisperer Modes
Feature matrix for LLMWhisperer modes
| Native Text | Low Cost | High Quality | Form | |
|---|---|---|---|---|
| PDF (not scanned) | ✓ | ✓ | ✓ | ✓ |
| PDF (scanned) | ✗ | ✓ | ✓ | ✓ |
| PDF (with forms) | ✗ | ✗ | ✗ | ✓ |
| Images | ✗ | ✓ | ✓ | ✓ |
| MS Office Document | ✗ | ✓ | ✓ | ✓ |
| MS Office Excel | ✗ | ✓ | ✓ | ✓ |
| MS Office Powerpoint | ✗ | ✓ | ✓ | ✓ |
| LibreOffice Writer | ✗ | ✓ | ✓ | ✓ |
| LibreOffice Calc | ✗ | ✓ | ✓ | ✓ |
| LibreOffice Impress | ✗ | ✓ | ✓ | ✓ |
| Checkbox and Radio button detection | ✗ No | ✗ No | ✗ No | ✓ Yes |
| Lines reproduction in output | ✗ No | ✓ Yes | ✓ Yes | ✓ Yes |
| Extraction performance | Very fast | Fast | Medium | Medium |
| Image preprocessing (median filter and gaussian blur) | ✗ No | ✓ Yes | ✗ No | ✗ No |
| Line splitting stratergy choice | ✓ Yes | ✓ Yes | ✓ Yes | ✓ Yes |
| Supported languages | All (unicode) | 120*+ | 300+ | 300+ |
| Handwritting recognition | ✗ No | Basic support | ✓ Yes | ✓ Yes |
| Layout preserving output | ✓ Yes | ✓ Yes | ✓ Yes | ✓ Yes |
| AI/ML based enhancement | ✗ No | ✗ No | ✓ Yes | ✓ Yes |
| Rotation and skew compensation | NA | ✗ No | ✓ Yes | ✓ Yes |
| Auto repair PDFs | ✓ Yes | ✓ Yes | ✓ Yes | ✓ Yes |
| Dense text content | Best performance | Very good | Very good | Very good |
| High entropy content (each page contains large variery of text sizes) | Best performance | Very good | Very good | Very good |
Recommended use cases
| Native Text | Low Cost | High Quality | Form | |
|---|---|---|---|---|
| Recommended use cases | • Low latency requirement • All documents are PDFs • PDFs are native text PDFs • Cost sensitive application | • High quality scanned PDFs • High quality scanned images • No handwritten documents | • Medium/low quality scanned PDFs • Medium/low quality scanned images • Handwritten documents | • Checkbox and radio button detection • Medium/low quality scanned PDFs • Medium/low quality scanned images • Handwritten documents |