Parse & Extract

Parse a document with the Parse API, then pull specific fields from the parsed Markdown with the Extract API. Extract takes Markdown and a JSON schema, so every extraction starts with a parse. The tutorial uses the ade-python library and ade-typescript library to:

Parse this PDF: Wire Transfer Form
Extract these fields: Bank Name and Total Invoice Amount
Save the parse response and the extract response as JSON files

These examples require the Python or TypeScript client library. Before running a script, set your API key and install the library and any required dependencies.

The scripts parse with a synchronous request, which consumes credits at the full rate. To parse many documents, submit them as Parse Jobs on the standard tier, which consume half the credits. See Parse a Directory of Documents.

1. Download the Document to Process

Download the Wire Transfer Form and save it to a local directory.

2. Create the Script

Copy the script for your language and save it as parse-extract.py or parse-extract.ts in the same directory as the PDF.

import json
from pathlib import Path
from landingai_ade import LandingAIADE

# Initialize client (uses VISION_AGENT_API_KEY environment variable)
client = LandingAIADE()

# Define the extraction schema
schema = {
    "type": "object",
    "properties": {
        "bank_name": {
            "description": "The official name of the bank where the account is held.",
            "x-alternativeNames": ["Name of Bank", "Financial Institution", "Bank"],
            "type": "string"
        },
        "total_invoice_amount": {
            "description": "The total monetary amount of the invoice, including all charges and taxes.",
            "x-alternativeNames": ["Grand Total", "Amount Due", "Invoice Total"],
            "type": "number"
        }
    }
}

# Parse the document
# save_to is optional. It saves the full parse response (Markdown, structure, and metadata)
# to output/wire-transfer_parse_output.json, which you need to ground the extracted values later
parse_response = client.v2.parse(
    document=Path("wire-transfer.pdf"),
    model="dpt-3-pro-latest",
    save_to="output"
)

# Extract fields from the parsed Markdown and save the extract response
extract_response = client.v2.extract(
    schema=schema,
    markdown=parse_response.markdown,
    model="extract-latest",
    save_to="output/wire-transfer_extract_output.json"
)

print(json.dumps(extract_response.extraction, indent=2))

import LandingAIADE from "landingai-ade";
import fs from "fs";

// Initialize client (uses VISION_AGENT_API_KEY environment variable)
const client = new LandingAIADE();

// Define the extraction schema
const schema = {
  type: "object",
  properties: {
    bank_name: {
      description: "The official name of the bank where the account is held.",
      "x-alternativeNames": ["Name of Bank", "Financial Institution", "Bank"],
      type: "string"
    },
    total_invoice_amount: {
      description: "The total monetary amount of the invoice, including all charges and taxes.",
      "x-alternativeNames": ["Grand Total", "Amount Due", "Invoice Total"],
      type: "number"
    }
  }
};

async function main(): Promise<void> {
  // Parse the document
  // saveTo is optional. It saves the full parse response (Markdown, structure, and metadata)
  // to output/wire-transfer_parse_output.json, which you need to ground the extracted values later
  const parseResponse = await client.v2.parse({
    document: fs.createReadStream("wire-transfer.pdf"),
    model: "dpt-3-pro-latest",
    saveTo: "output"
  });

  // Extract fields from the parsed Markdown and save the extract response
  const extractResponse = await client.v2.extract({
    schema: schema,
    markdown: parseResponse.markdown,
    model: "extract-latest",
    saveTo: "output/wire-transfer_extract_output.json"
  });

  console.log(JSON.stringify(extractResponse.extraction, null, 2));
}

main();

3. Run the Script

Run the script from the same directory:

python parse-extract.py

npx tsx parse-extract.ts

4. View the Output

The script prints the extracted values and saves two files to an output folder in the same directory:

File	Contents
`wire-transfer_parse_output.json`	The full parse response: `markdown`, the `structure` tree of pages and blocks with the page number and bounding box of each, and `metadata`. See Parse API Response.
`wire-transfer_extract_output.json`	The full extract response. See Extract API Response.

The extract response has these top-level fields:

Field	Contents
`extraction`	The extracted values, in the shape of your schema: `bank_name` and `total_invoice_amount`.
`extraction_metadata`	The same shape, with each value replaced by a `{value, ranges}` object. Each range is a `{start, end}` pair of character offsets into the Markdown the value was copied from, and `ranges` is `null` for a value the model did not copy from the document. A value that appears in more than one place has one range per place, like `total_invoice_amount` on the wire transfer form.
`markdown`	The Markdown you sent, echoed back unchanged. Every range indexes into this string.
`metadata`	Request details, including `doc_id`: the `job_id` of the parse whose Markdown was extracted.
`schema_violation_error` and `warnings`	Set when the extraction is partial or has non-fatal warnings. See Response Structure.

A run on the wire transfer form saved this wire-transfer_extract_output.json, shown with the IDs and most of the Markdown shortened. Your ranges and metadata can differ, because they depend on your parse and the model version:

wire-transfer_extract_output.json

{
  "extraction": {
    "bank_name": "JPMorgan Chase Bank, N.A.",
    "total_invoice_amount": 15750.0
  },
  "extraction_metadata": {
    "bank_name": {
      "value": "JPMorgan Chase Bank, N.A.",
      "ranges": [{ "start": 186, "end": 224 }]
    },
    "total_invoice_amount": {
      "value": 15750.0,
      "ranges": [{ "start": 960, "end": 970 }, { "start": 112, "end": 150 }]
    }
  },
  "markdown": "WIRE TRANSFER FORM\n\n\n### Invoice Information\n\n...",
  "metadata": {
    "job_id": "extract-...",
    "doc_id": "parse-...",
    "model_version": "extract-20260819",
    "billing": { "service_tier": "priority", "total_credits": 1.0 },
    "range_units": "unicode_codepoints"
  },
  "schema_violation_error": null,
  "warnings": []
}

Next Steps

Now that you have a working script, you can:

Replace wire-transfer.pdf with any document you want to parse and extract from.
Modify the schema to extract different fields. For guidance, see Extraction Schema (JSON).
Find each extracted value on the page: pass extraction_metadata and the parse’s structure to the Ground API. See Link Extracted Data to Document Locations.

Overview

Parse Only

Parse & Extract

1. Download the Document to Process

2. Create the Script

3. Run the Script

4. View the Output

Next Steps

​1. Download the Document to Process

​2. Create the Script

​3. Run the Script

​4. View the Output

​Next Steps

1. Download the Document to Process

2. Create the Script

3. Run the Script

4. View the Output

Next Steps