Docupie is an advanced document processing tool that leverages AI to extract structured data from PDFs. It is built to handle PDF conversions, extract relevant information, and format results as specified by customizable schemas.

• Extracts structured JSON output from unstructured documents.
• Converts documents into Markdown format.
• Supports custom schemas for data extraction.
• Includes pre-defined templates for common schemas.
• Works with OpenAI and custom LLM setups (Llava and Llama3.2-vision).
• Auto-generates schemas based on document content.

The hosted version provides a seamless experience with fully managed APIs, so you can skip the setup and start extracting data right away. Join the beta to get access to the hosted service.

In the meantime, you can explore the playground here. Upload your documents and extract structured data with your own custom schema, or use one of the sample documents and template schemas.

• PDF Extraction
• Basic Schema Definition
• Structured JSON Output
• Template Schemas
• Local LLM Integration (Llava and Llama3.2)
• Auto-generated Schemas
• Documnt Formatters (Text and Markdown)
• Multi-file Support (DOCX, PNG, JPG, TXT, HTML)
• Additional Schema Field Types (Boolean and Enum)

• Extended LLM Support (Local and cloud)
• Image Data Extraction
• Advanced Document Formatters
• Data Classification

Before using Docupie, ensure the following dependencies are installed:

• Ghostscript: Docupie relies on Ghostscript for handling certain PDF operations.
• GraphicsMagick: Required for image processing within document conversions.

Install both on your system before proceeding:

# On macOS brew install ghostscript graphicsmagick # On Debian/Ubuntu sudo apt-get update sudo apt-get install -y ghostscript graphicsmagick

Ensure Python 3.10.4+ is installed on your system.

You can install Docupie via pip:

pip install Docupie

Docupie requires an .env file to store sensitive information like your OpenAI API key.

Create an .env file in your project directory and add the following:

OPENAI_API_KEY=your_openai_api_key

First, import Docupie and define your schema. The schema outlines what information Docupie should look for in each document. Here's a quick setup to get started.

The schema is a list of dictionaries where each dictionary defines:

• name: Field name to extract.
• type: Data type (e.g., "string", "number", "array", "object").
• description: Description of the field.
• children (optional): For arrays and objects, define nested fields.

Example schema for a bank statement:

schema = [ { "name": "accountNumber", "type": "string", "description": "The account number of the bank statement." }, { "name": "openingBalance", "type": "number", "description": "The opening balance of the account." }, { "name": "transactions", "type": "array", "description": "List of transactions in the account.", "children": [ { "name": "date", "type": "string", "description": "Transaction date." }, { "name": "creditAmount", "type": "number", "description": "Credit Amount of the transaction." }, { "name": "debitAmount", "type": "number", "description": "Debit Amount of the transaction." }, { "name": "description", "type": "string", "description": "Transaction description." } ] }, { "name": "closingBalance", "type": "number", "description": "The closing balance of the account." } ]

Use Docupie to process a PDF by passing the file URL and the schema.

from docupie import extract async def run_extraction(): result = await extract( file="https://bank_statement.pdf", schema=schema ) print("Extracted Data:", result) # If using asyncio import asyncio asyncio.run(run_extraction())

Here’s an example of what the extracted result might look like:

 { "success": true, "pages": 1, "data": { "accountNumber": "100002345", "openingBalance": 3200, "transactions": [ { "date": "2021-05-12", "creditAmount": null, "debitAmount": 100, "description": "transfer to Tom" }, { "date": "2021-05-12", "creditAmount": 50, "debitAmount": null, "description": "For lunch the other day" }, { "date": "2021-05-13", "creditAmount": 20, "debitAmount": null, "description": "Refund for voucher" }, { "date": "2021-05-13", "creditAmount": null, "debitAmount": 750, "description": "May's rent" } ], "closingBalance": 2420 }, "fileName": "bank_statement.pdf" } 

Read the documentation for more on how to define schemas and and enable auto-generation.

Docupie comes with built-in templates for extracting data from popular document types like invoices, bank statements, and more. These templates make it easier to get started without defining your own schema.

List available templates

You can list all available templates using the list_templates function.

from Docupie import templates available_templates = templates.list() print(available_templates) # Prints all available template names

Use a template

To use a template, simply pass its name to the extract function along with the file you want to extract data from. Here's an example:

from Docupie import extract import asyncio async def run_extraction(): result = await extract( file="https://bank_statement.pdf", template="bank_statement" ) print("Extracted Data:", result) asyncio.run(run_extraction())

Read the templates documentation for more details on templates and how to contribute yours.

Read more on how to use local models here.

Contributions are welcome! Please submit a pull request with any improvements or features.

This project is licensed under the AGPL v3.0 License.

This project is a Python port of the Documind package. We extend our gratitude to the Documind team for their work, which served as the foundation for Docupie. This project is published under the AGPLv3 license as defined in the LICENSE file.

This repo was also built on top of Zerox. The MIT license from Zerox is included in the core folder and is also mentioned in the root license file.