Quickstart

While Retab is built around evals, the most basic workflow is extracting structured data from a document. The easiest way to access the API is through the Python or Node SDK.

Find your API Key in the dashboard settings.

Using the Retab SDK

1. Install the SDK

   Python

   pip install retab
   
2. Generate a Schema

   Python

   from retab import Retab
   
   client = Retab()
   
   schema_response = client.schemas.generate(
       documents=["Invoice.pdf"],
       model="retab-small",
   )
   
3. Extract Data

   Python

   from retab import Retab
   
   client = Retab()
   
   extraction_response = client.documents.extract(
       json_schema = schema_response.json_schema,
       document = "Invoice.pdf",
       model="retab-micro",
   )
   
   print(extraction_response.data)
   print(extraction_response.likelihoods)
   print(extraction_response.extraction_id)
   

Congratulations! You’ve extracted structured data from a document with Retab. Now you can start building evals, deploy processors, automate workflows and more.