Welcome to Sensible's documentation. Sensible extracts structured data from documents, for example, business forms in PDF format.
Sensible uses an extraction configuration, or "config", to extract data from a collection of similar documents that you send to Sensible, either through our app or through our API. You have the following options for authoring configs:
|Use case||Description||Get started|
|Low-code AI authoring||You describe the document data that you want to extract, using Sensible's visual authoring tool, Sensible Instruct. Sensible uses GPT-4 and other large-language models (LLMs) to extract data from your documents.||Getting started with AI extractions|
|Advanced authoring||To extract from complex document layouts, use SenseML, a superset of Sensible Instruct. SenseML is a JSON-formatted query language that combines layout-based queries with AI-powered queries. This combined approach lets you extract from a variety of documents, from highly structured business forms to unstructured legal or research papers.||Getting started with SenseML|
|Out-of-the-box extraction for common business documents||No need to author configs for common business and tax forms. Use Sensible's pre-built, open-source configuration library to extract key information from tax forms such as 1099s, major carrier insurance declaration pages, and other documents. Then tweak the pre-built configurations for your custom data needs.||Getting started with out-of-the-box extractions|
For more information about choosing between SenseML and Sensible Instruct, see Choosing an extraction approach.
See the following topics for further resources:
Updated about 10 hours ago