documind Documentation

A schema is an array of objects that defines the structure of the data you want to extract. Each object must include:

• name: The field name to extract.
• type: The data type (e.g., "string", "number", "array", "object", "boolean", or "enum").
• description: A description of the field to guide the AI.
• children (optional): An array of nested field objects, used for "array" or "object" types.

const schema = [
  {
    name: "accountNumber",
    type: "string",
    description: "The account number of the bank statement."
  },
  {
    name: "transactions",
    type: "array",
    description: "List of transactions in the account.",
    children: [
      {
        name: "date",
        type: "string",
        description: "Transaction date."
      }
    ]
  }
];

Install the documind package via npm. Ensure you have Node.js (v18+) and NPM installed on your system.

npm install documind

Documind requires Ghostscript (for PDF operations) and GraphicsMagick (for image processing) to be installed on your operating system.

On macOS:

brew install ghostscript graphicsmagick

On Debian/Ubuntu:

sudo apt-get update
sudo apt-get install -y ghostscript graphicsmagick

# On macOS
brew install ghostscript graphicsmagick

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

Documind provides pre-defined templates for common documents like invoices and bank statements.

1. List available templates using templates.list().
2. Apply a template by passing the template name to the template property in the extract() options object.

Documind requires an .env file in your project directory to store your OpenAI API key. Create the file and add the following variable:

OPENAI_API_KEY=your_openai_api_key

Use the extract function to process a document. You can provide either a custom schema or a pre-defined template name. The file parameter accepts a URL to the document.

import { extract } from 'documind';

const runExtraction = async () => {
  const result = await extract({
    file: 'https://bank_statement.pdf',
    schema
  });

  console.log("Extracted Data:", result);
};

runExtraction();

The Ollama provider requires the following environment variable to be set:

• BASE_URL: The base URL of your local Ollama instance (e.g., http://localhost:11434).

If BASE_URL is not provided, the provider will throw an error: Missing BASE_URL in environment variables.

The Google provider requires a valid API key to authenticate requests to the Google Generative Language API. This must be provided via the GEMINI_API_KEY environment variable.

If this variable is missing, the provider will throw an error: Missing GEMINI_API_KEY in environment variables.

The OpenAI provider requires an API key to authenticate requests to OpenAI. This must be provided via the following environment variable:

• OPENAI_API_KEY: Your OpenAI API key.

The extract function returns a JSON object containing the extraction status, page count, the extracted data, and the filename.

Example Output Structure:

 {
  "success": true,
  "pages": 1,
  "data": {
    "accountNumber": "100002345",
    "openingBalance": 3200,
    "transactions": [
        {
        "date": "2021-05-12",
        "creditAmount": null,
        "debitAmount": 100,
        "description": "transfer to Tom" 
      }
    ],
    "closingBalance": 2420
  },
  "fileName": "bank_statement.pdf"
}