Extractions are at the heart of Flora’s AI-powered data processing capabilities. This endpoint allows you to convert unstructured information from various sources into structured, actionable data based on predefined schemas.

​
Key Features

  • Multi-Source Support: Extract data from images, PDFs, text, and webpages.
  • Schema-Based Extraction: Utilize custom schemas to define the structure of extracted data.
  • Flexible AI Models: Choose between different AI models for optimal performance.

​
Supported Types and Sources

Flora supports the following types of unstructured data:
  • Image: Provide a valid and accessible image URL.
  • PDF: Supply a valid and accessible PDF URL.
  • Text: Directly input the text content to extract data from.
  • Webpage: Provide a valid and accessible webpage URL.
Flora only supports processing unstructured data via URLs and does not provide file upload capabilities. Ensure your source files are hosted and accessible via a public URL.

​
Workflow

  1. Create a Schema: Define the structure for your extracted data using the Schemas endpoint.
  2. Initiate Extraction: Send a request to the Create Extraction endpoint with your source URL and schema.
  3. Retrieve Results: Once completed, access your structured data from the extraction response.

​
Best Practices

  • Ensure your schemas are well-defined to maximize extraction accuracy.
  • Utilize the optional instruction parameter to guide the AI in complex extractions.
  • Verify that your source URLs are publicly accessible and valid before submitting an extraction request.
Leverage the power of Flora’s Extractions API to unlock valuable insights from your unstructured data sources.