What types of documents does Unsiloed AI support?
Unsiloed AI processes a wide range of formats including PDFs, images, spreadsheets, and scanned documents. It can handle mixed layouts such as tables, charts, forms, and handwritten content within a single file, making it versatile for various enterprise data needs.
How is Unsiloed AI different from traditional OCR?
Traditional OCR provides a flat stream of text. Unsiloed AI uses advanced vision models to understand document structure, preserving tables, sections, and parent-child relationships. This ensures the output is truly LLM-ready and contextually rich, unlike basic text extraction.
What kind of output does Unsiloed AI generate?
Unsiloed AI generates clean, LLM-ready Markdown and structured JSON, complete with confidence scores for each field. For extraction tasks requiring a guaranteed shape, fully schema-validated outputs are also available, ensuring data integrity and usability.
Can Unsiloed AI be deployed in a private VPC or on-premise?
Yes, Unsiloed AI supports both managed and air-gapped deployments. The API and outputs remain consistent across both modes, ensuring seamless integration regardless of your infrastructure requirements for security and data residency.
What is included in the free Builder plan?
The free Builder plan includes support for 2,000 pages, PDF and image file support, layout extraction, table extraction, form field extraction, table summarization, and image summarization. It's a great way to start experimenting with the tool's capabilities.