AP News Data API: Extract Structured JSON in 2026

This guide covers extracting publicly accessible data. Always review a site's robots.txt and Terms of Service before scraping.

TL;DR

To get structured AP News data via API, use AlterLab's Extract API with a defined JSON schema for the fields you need (headline, author, etc.). Simply POST the article URL and schema to receive validated, typed JSON output — no HTML parsing required. This approach handles anti-bot measures and delivers ready-to-use data for pipelines.

Why use AP News data?

AP News provides timely, authoritative coverage across global events, business, and technology. Extracting structured data enables:

• AI training: Feed NLP models with real headlines and summaries for sentiment analysis or topic modeling.
• Analytics: Track keyword frequency or entity mentions over time to identify emerging trends.
• Content enrichment: Augment internal datasets with verified news metadata for recommendation systems.

What data can you extract?

From publicly accessible AP News article pages, you can reliably extract:

• headline: The main title of the article
• author: Byline or reporting entity
• published_date: ISO timestamp of publication
• source: Typically "Associated Press" or wire service credit
• summary: Lead paragraph or abstract These fields appear consistently in article metadata and are safe to extract as public information.

The extraction approach

Direct HTTP requests to apnews.com return HTML that requires fragile parsing. Selectors break when AP News updates its frontend, and anti-bot measures (like JavaScript challenges) necessitate headless browsers — adding maintenance overhead. AlterLab's data API solves this by:

• Automatically handling JavaScript rendering and proxy rotation
• Applying your JSON schema to extract only specified fields
• Returning typed, validated output ready for downstream consumption This shifts the burden from HTML wrangling to data utilization.

Quick start with AlterLab Extract API

Begin by installing the AlterLab SDK (getting started guide). The Extract API converts a URL and schema into structured JSON.

import alterlab

client = alterlab.Client("YOUR_API_KEY")

schema = {
  "type": "object",
  "properties": {
    "headline": {
      "type": "string",
      "description": "The headline field"
    },
    "author": {
      "type": "string",
      "description": "The author field"
    },
    "published_date": {
      "type": "string",
      "description": "The published date field"
    },
    "source": {
      "type": "string",
      "description": "The source field"
    },
    "summary": {
      "type": "string",
      "description": "The summary field"
    }
  }
}

result = client.extract(
    url="https://apnews.com/example-article",
    schema=schema,
)
print(result.data)

Equivalent cURL request:

curl -X POST https://api.alterlab.io/v1/extract \
  -H "X-API-Key: YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "url": "https://apnews.com/example-article",
    "schema": {"properties": {"headline": {"type": "string"}, "author": {"type": "string"}, "published_date": {"type": "string"}}}
  }'

For batch processing, use async jobs:

import asyncio
import alterlab

async def extract_article(url, schema):
    client = alterlab.Client("YOUR_API_KEY")
    return await client.extract_async(url=url, schema=schema)

async def main():
    urls = [
        "https://apnews.com/article1",
        "https://apnews.com/article2",
        "https://apnews.com/article3"
    ]
    schema = {"type": "object", "properties": {"headline": {"type": "string"}}}
    tasks = [extract_article(url, schema) for url in urls]
    results = await asyncio.gather(*tasks)
    for r in results:
        print(r.data)

if __name__ == "__main__":
    asyncio.run(main())
"

Define your schema

The JSON schema parameter tells AlterLab exactly which fields to extract and their expected types. In the Python example above:

• Each property specifies type: string (AlterLab also supports numbers, booleans, arrays)
• The description field aids model understanding but isn't required for validation
• AlterLab validates the LLM's output against this schema before returning data If the extracted data doesn't match (e.g., a number where a string is expected), the API returns a validation error — preventing malformed data from entering your pipeline.

Handle pagination and scale

For high-volume extraction:

1. Batching: Group 10-50 URLs per async job to optimize throughput
2. Rate limiting: AlterLab enforces polite defaults; adjust concurrency based on your /pricing tier
3. Error handling: Implement retries with exponential backoff for transient 429s
4. Cost estimation: Use the /v1/estimate endpoint to preview expenses before large batches Remember: AlterLab manages infrastructure scaling, but you remain responsible for respecting AP News's crawl-delay in robots.txt and implementing application-level throttling.

Key takeaways

• Structured output via schema eliminates brittle HTML parsing
• Validation ensures data consistency for ML models and analytics
• AlterLab handles compliance complexity for public data extraction
• Cost scales linearly with usage — see pricing for details
• Start small: extract single articles before scaling to pipelines