Skip to main content
Projects provide a systematic way to test and validate your extraction schemas against known ground truth data. Think of it as evals for document AI. You can measure accuracy, compare different models, and optimize your extraction pipelines with confidence. A project consists of documents with annotations (your test data), iterations (test runs with different settings), and a schema (what you want to extract). This structure lets you run A/B tests between models and systematically improve your document processing accuracy.

​
Deployments

Deployments are project-based configurations for document extraction that can be called via the API route https://api.retab.com/v1/projects/extract/{project_id}. This is the primary method for executing document extraction using project-based configurations.
​
Returns
RetabParsedChatCompletion
The extracted data as a JSON object matching the project’s schema.
from retab import Retab

client = Retab()

completion = client.projects.extract(
    project_id="proj_01G34H8J2K",
    document="invoice.pdf"
)

​
Parameters

​
project_id
string
required
ID of the project
​
iteration_id
string
required
ID of the specific iteration to use, or "base-configuration" to use the project’s default settings.
​
document
Path | str | bytes | IOBase | MIMEData | PIL.Image.Image | HttpUrl
Document to process.
​
metadata
dict[str, str]
Custom key-value metadata to attach to the extraction for organization and filtering.

​
Organizing Extractions with Metadata

The metadata parameter allows you to attach custom key-value pairs to your extractions, making it easy to organize, filter, and retrieve results later. This is particularly useful when processing documents for multiple clients, departments, or workflows within a single project. 
from retab import Retab

client = Retab()

# Attach metadata to organize extractions by customer
completion = client.projects.extract(
    project_id="proj_01G34H8J2K",
    document="invoice.pdf",
    metadata={
        "organization_id": "org_acme_corp",
        "department": "accounting",
        "batch_id": "batch_2024_04"
    }
)
Common use cases for metadata include:
  • Multi-tenant applications: Tag extractions with organization_id or customer_id to separate data by client
  • Workflow tracking: Add batch_id, pipeline_stage, or source_system to trace document processing
  • Categorization: Include document_type, region, or priority for filtering and reporting
Please check the API Reference for complete method documentation.