LLMWhisperer Cloud Edition
LLMWhisperer Cloud is a fully managed SaaS offering that makes complex document data ready for LLM consumption. No infrastructure to manage — sign up, get an API key, and start extracting text from your documents.
Key Features
- Layout-preserving text extraction for maximum accuracy from LLMs
- Checkbox and radio button detection in common forms, presented in an LLM-friendly way
- Auto-compaction to reduce LLM token consumption while preserving layout
- Image preprocessing (Gaussian Blur, Median Filtering, Contrast Adjustments) to improve text clarity in scanned documents
- 300+ language support with handwriting recognition
- Auto mode switching from native text to OCR as needed
- Auto PDF repair for corrupted or malformed documents
- Support for PDFs, images, MS Office, LibreOffice, and more
Processing Modes
LLMWhisperer Cloud supports multiple processing modes, each optimized for different document types and use cases:
| Mode | Best For |
|---|---|
native_text | Digital PDFs with embedded text |
low_cost | Cost-effective OCR for scanned documents |
high_quality | AI/ML-enhanced OCR with handwriting support |
form | Documents with checkboxes, radio buttons, and form fields |
table | Documents with complex tabular structures |
See LLMWhisperer Modes for the full feature matrix and recommended use cases.
API Regions
LLMWhisperer Cloud is available in the following regions:
| Region | Base URL |
|---|---|
| US Central | https://llmwhisperer-api.us-central.unstract.com/api/v2 |
| EU West | https://llmwhisperer-api.eu-west.unstract.com/api/v2 |
Pricing
- Pay-as-you-go — no flat fee or upfront commitment
- 100 free pages daily at no cost
- Tiered pricing from $1 to $15 per 1,000 pages depending on mode
- Metered based on the number of pages processed per mode
Visit the Pricing page for detailed pricing information.
Getting Started
- Register for an LLMWhisperer account
- Try the Playground to test document extraction
- Integrate using the API Reference, Python Client, or JavaScript Client