Global News Archive API - Rise of the Phoenix

Search a hosted global news archive by country, publisher, and date, or pull the persisted daily country/topic briefs generated by the scraper's LLM summary pipeline.

Rise of the Phoenix is a historical news API Actor for researchers, analysts, media teams, and AI builders who need article data or ready-to-use daily intelligence briefs without running a live crawl every time.

Because it searches a hosted archive, runs are fast, predictable, easy to paginate, and suitable for repeatable monitoring jobs or large historical exports.

Use it when you need a news API for:

• Media monitoring by country, publisher, or date range.
• Historical article exports for research, PR, reputation monitoring, or market intelligence.
• Clean news text datasets for RAG, LLM evaluation, NLP, classification, and summarization workflows.
• A fast first pass before deciding whether a broader live crawl is worth running.
• Country-level daily topics with grounded source references, evidence counts, and importance scores.

Why Use This Actor

• Archive-first workflow: query stored article records instead of crawling pages live.
• Useful filters: narrow exports by country, publisher, and publication date.
• Estimate before export: check available volume before paying for a larger dataset.
• Pagination built in: continue broad exports with nextContinuationToken.
• Clean outputs: get article title, body, URL, source, country, language, image, and metadata fields in the Apify dataset.
• Daily LLM briefs: retrieve summaries already generated and stored by the scraper pipeline; the Actor does not make a new LLM request.

Quick Starts

Monitor Recent AP News Articles

{
  "websites": ["AP News"],
  "published_from": "10 days",
  "published_to": "0 days",
  "max_results": 25
}

Build A Country-Level News Dataset

{
  "countries": ["United States"],
  "published_from": "30 days",
  "published_to": "0 days",
  "max_results": 100
}

Pull Daily LLM Summaries

{
  "output_mode": "daily_summaries",
  "countries": ["South Africa", "Ghana"],
  "published_from": "7 days",
  "published_to": "0 days",
  "max_results": 100
}

Check Archive Volume Before Exporting

{
  "countries": ["South Africa"],
  "published_from": "12 months",
  "published_to": "0 days",
  "estimate_only": true,
  "max_results": 100
}

Media Monitoring For A Market

{
  "countries": ["United Kingdom"],
  "published_from": "7 days",
  "published_to": "0 days",
  "max_results": 100
}

Publisher-Specific Research

{
  "websites": ["Reuters", "AP News"],
  "published_from": "30 days",
  "published_to": "0 days",
  "max_results": 100
}

What You Can Do

• Search historical news articles by country or website.
• Pull persisted daily country/topic LLM summaries by country and date.
• Filter article data by publication date range.
• Estimate result counts before exporting a large dataset.
• Export article records to the Apify dataset in JSON, CSV, Excel, XML, RSS, or HTML.
• Continue large exports with cursor-based pagination.
• Use the results for media monitoring, news research, market intelligence, academic research, lead enrichment, and AI or LLM dataset preparation.

Common Use Cases

• Media monitoring: track recent coverage in a country or across selected publishers.
• PR and reputation research: export historical mentions and articles for review.
• Market intelligence: collect regional news context for companies, sectors, products, and public events.
• AI and RAG datasets: feed clean article text into analytics, NLP, summarization, classification, retrieval, and evaluation pipelines.
• Research scoping: estimate how much article data is available before launching a larger export.

Best For

• Users who need structured article records from an archive.
• Analysts who need generated daily briefs with links back to supporting articles.
• Teams that want repeatable exports from the same source and date filters.
• Workflows where result count estimates and pagination matter.
• API users who want JSON, CSV, Excel, XML, RSS, or HTML exports from Apify datasets.

Not For

• Real-time breaking-news crawling from arbitrary websites.
• Scraping every page from a website outside the hosted archive.
• Bypassing publisher restrictions, paywalls, or access controls.

Troubleshooting

• If a country or publisher returns no rows, widen the date range or check that the selected country/source exists in the Actor input list.
• If a large export stops before all available rows are returned, use nextContinuationToken from OUTPUT to continue.
• If you only need a volume check, enable estimate_only before exporting article rows.
• If you need current breaking-news crawling from arbitrary websites, use a live crawler instead of this archive API.

Pricing and Cost Notes

• Keep public example tasks between 25 and 100 results so they are quick and affordable.
• Use estimate_only to scope large country/date searches before exporting.
• Increase max_results only when you need larger exports; broad searches can return many article rows.

How To Run The Actor

1. Open the Actor on Apify.
2. Choose Archived articles or Daily LLM summaries in Output mode.
3. For articles, leave AP News selected for a quick test or choose Countries/Websites. For daily summaries, optionally choose Countries; no country selection returns all countries in the bounded date window.
4. Set Date from and Date to.
5. Optional: enable Advanced: estimate results first to get a count without returning rows.
6. Set Max results, run the Actor, then download the Dataset and read pagination details from the OUTPUT record.

Input

The Actor input form contains the full list of available countries and websites. The main fields are:

Field	Type	Required	Description
output_mode	string	No	articles (default) or daily_summaries.
countries	string[]	No	Search one or more countries. Use this or websites, not both.
websites	string[]	No	Search publishers in article mode. Daily summaries are country-level and do not accept website filters.
published_from	string	No	Article publication start or summary-date start. Supports ISO dates and relative values.
published_to	string	No	Article publication end or summary-date end. Supports ISO dates and relative values.
max_results	integer	No	Maximum records to return. Default is 10; maximum is 5000.

Advanced fields:

Field	Type	Required	Description
estimate_only	boolean	No	Estimate matching records without returning dataset rows. Works in both modes.
continuation_token	string	No	Token from a previous run in the same output mode.

Helpful defaults:

• AP News is selected by default. If you clear all countries and websites, the Actor still searches AP News.
• In daily_summaries mode, the implicit AP News default is removed. An empty country selection returns all countries, still bounded by the date window and max_results.
• If you provide no dates, the Actor searches the last 10 days.
• If you provide only published_from, published_to defaults to 0 days.
• If you provide only published_to, the Actor derives a one-day window ending at that date.
• Duplicate countries or websites are ignored automatically.
• Invalid or cross-mode continuation tokens fail fast before the Actor queries storage.

Accepted date examples:

• 2025-01-01
• 2025-01-01T00:00:00Z
• 7 days
• 30 days
• 12 months
• 2 years

Example Inputs

Search By Country

{
  "countries": ["South Africa"],
  "published_from": "30 days",
  "published_to": "0 days",
  "max_results": 100
}

Search By Website

{
  "websites": ["AP News", "Reuters"],
  "published_from": "2025-01-01",
  "published_to": "2025-12-31",
  "max_results": 100
}

Estimate Results Before Exporting

{
  "countries": ["United States"],
  "published_from": "12 months",
  "published_to": "0 days",
  "estimate_only": true,
  "max_results": 500
}

Continue A Large Export

{
  "countries": ["Africa"],
  "published_from": "2 years",
  "published_to": "0 days",
  "max_results": 100,
  "continuation_token": "{\"date_published\":\"2026-05-06T15:44:00+00:00\",\"url_hash\":\"81a489c65af24950956dd717c2f7b4be\"}"
}

Pull Daily Summaries For A Date Range

{
  "output_mode": "daily_summaries",
  "countries": ["South Africa"],
  "published_from": "2026-07-01",
  "published_to": "2026-07-22",
  "max_results": 100
}

Output Data

When estimate_only is false, the selected record type is pushed to the default Apify dataset while the run is in progress. If a run stops early, already pushed records remain available. Every row has record_type set to article or daily_summary.

Archived Article Fields

Field	Type	Description
site_name	string	News website or publisher name.
country	string	Country associated with the source.
region	string	Broader region associated with the source.
language	string	Source language metadata.
article_title	string	Article headline.
author	string | null	Author or byline when available.
article_body	string	Normalized article text.
tags	string[]	Tags or keywords when available.
date_published	string	ISO 8601 publication timestamp.
article_url	string	Canonical article URL.
main_image_url	string | null	Featured image URL when available.
seo_description	string | null	Article summary or meta description when available.
execution_mode	string	current for incremental scrapes or historic for archive backfills.
summary_available	boolean	Whether processed LLM enrichment is attached.
short_summary	string | null	Concise grounded article summary.
event_summary	string | null	Summary of the real-world event.
bullet_summary	string[]	Three concise summary bullets when enrichment is available.
main_topic	string | null	Normalized primary topic.
event_type	string | null	Normalized event classification.
entities	object	Extracted people, companies, organisations, and locations.
keywords	string[]	Grounded keywords extracted during enrichment.
summary_confidence	number | null	Enrichment confidence from 0 to 1.

Example dataset item:

{
  "site_name": "AP News",
  "country": "United States",
  "region": "North America",
  "language": "en",
  "article_title": "Sample archived news headline",
  "author": null,
  "article_body": "Normalized article text appears here...",
  "tags": [],
  "date_published": "2026-05-06T16:13:00+00:00",
  "article_url": "https://example.com/news/sample-article",
  "main_image_url": null,
  "seo_description": null,
  "execution_mode": "current",
  "summary_available": true,
  "main_topic": "energy",
  "short_summary": "A concise grounded summary of the article.",
  "event_summary": "The article reports a material development in the energy sector.",
  "bullet_summary": ["Development announced.", "Stakeholders are affected.", "Further action is expected."],
  "summary_confidence": 0.91
}

Daily LLM Summary Fields

Daily summary mode returns one stored country/topic row per dataset item. It does not call OpenRouter or another model during the Actor run.

Field	Type	Description
summary_date	string	UTC country-day represented by the brief.
country	string	Country covered by the topic.
topic_rank	integer	Topic order within the country/day brief.
topic_category	string	Normalized topic category.
topic_headline	string	Brief headline.
topic_summary	string	Grounded country/topic summary.
topic_importance_score	number	Calibrated topic importance score from 0 to 100.
key_points	string[]	Concise facts supporting the brief.
article_refs	object[]	Supporting article IDs, URLs, sources, and timestamps.
source_count	integer	Distinct sources supporting the topic.
article_count	integer	Articles linked to the topic.
input_article_count	integer	Qualifying daily articles considered for the country.
llm_model	string	Model that generated the stored brief.
generated_at	string	Generation timestamp.

Example daily summary item:

{
  "record_type": "daily_summary",
  "summary_date": "2026-07-22",
  "country": "South Africa",
  "topic_rank": 1,
  "topic_category": "Infrastructure & Energy",
  "topic_headline": "Grid investment plan advances",
  "topic_summary": "Officials advanced a national grid investment plan. Implementation details remain the key uncertainty.",
  "topic_importance_score": 78,
  "source_count": 2,
  "article_count": 3,
  "llm_model": "qwen/qwen3-8b"
}

Output Summary And Pagination

Each run also writes an OUTPUT record with summary and pagination metadata:

• resultCount
• summarizedResultCount in article mode
• hasMore
• nextContinuationToken
• filters
• estimatedMatchCount when estimate_only is enabled
• estimatedReturnedThisRun when estimate_only is enabled

If hasMore is true, run the Actor again with the same filters and pass nextContinuationToken into continuation_token.

In estimate mode:

• The dataset remains empty.
• The OUTPUT record includes the estimated match count.
• You can rerun the same input with estimate_only: false to fetch rows in the selected mode.

Python API Example

Use the Actor ID from the Actor's API tab.

import os
from apify_client import ApifyClient

client = ApifyClient(os.environ["APIFY_TOKEN"])

ACTOR_ID = "thescrapelab/Apify-The-Rise-of-the-Phoenix"

run_input = {
    "output_mode": "articles",
    "countries": ["South Africa"],
    "published_from": "30 days",
    "published_to": "0 days",
    "max_results": 25,
}

run = client.actor(ACTOR_ID).call(run_input=run_input)

dataset_items = list(
    client.dataset(run["defaultDatasetId"]).iterate_items(clean=True)
)

output_record = client.key_value_store(
    run["defaultKeyValueStoreId"]
).get_record("OUTPUT")
summary = output_record["value"] if output_record else {}

print("Run status:", run["status"])
print("Records returned:", len(dataset_items))
print("Has more:", summary.get("hasMore"))
print("Next cursor:", summary.get("nextContinuationToken"))

To check volume before exporting rows, set estimate_only to true. The Actor will return the estimate in the OUTPUT record and leave the dataset empty.

To pull daily briefs instead, use the same API flow with:

run_input = {
    "output_mode": "daily_summaries",
    "countries": ["South Africa"],
    "published_from": "7 days",
    "published_to": "0 days",
    "max_results": 100,
}

Tips For Better Results

• Use estimate_only before broad searches such as large countries, long date ranges, or global sources.
• Use narrower date ranges when you need smaller, faster exports.
• Use websites when you need publisher-specific news data.
• Use daily_summaries when you need stored country/day topics; no new LLM charge is incurred by retrieval.
• Keep the same filters when using continuation_token; only the token should change between pages.
• Increase max_results when you want fewer API calls, up to the Actor limit.

Pricing

This Actor uses pay-per-event pricing. You are charged a small Actor start event plus a per-result event for each dataset item returned. Estimate-only runs do not return dataset items, so they are useful for checking query size before exporting a large dataset.

Recommended usage:

• Run a small query first with max_results between 10 and 100.
• Use estimate_only for broad archive searches.
• Export larger pages with max_results up to 5000 when the estimate looks right.
• The measured default runtime is 128 MB memory with a 300-second timeout. The Actor is database-backed, so higher memory is rarely needed unless Apify support asks you to test it.

FAQ

Does this Actor scrape websites live during the run?

No. It reads matching articles or daily summaries from hosted storage. This makes it useful as a fast historical news API rather than a live crawling job.

Does daily-summary mode generate new summaries?

No. It pulls country/topic summaries already generated by the scraper's scheduled LLM pipeline. A missing country/day therefore returns no summary rows instead of generating them on demand.

Do I need to manage infrastructure to use it?

No. Run the Actor from Apify Console, tasks, schedules, or the Apify API. Provide your search filters and download the results from the dataset.

Why did I get zero dataset items?

The most common reasons are:

• Your filters matched no archived articles or stored daily summaries.
• Your date range is too narrow.
• estimate_only was enabled.
• Your continuation_token points beyond the available results.

How do I fetch more than one page of results?

Check the OUTPUT record. If hasMore is true, copy nextContinuationToken into the next run as continuation_token and keep the same output mode and filters.

What can I export?

You can export the dataset from Apify as JSON, CSV, Excel, XML, RSS, or HTML. Programmatic users can fetch the dataset through the Apify API.

Are results streamed during the run?

Yes. Result batches are written to the default dataset as they are read from the archive, so partial results remain available if a run is interrupted.

Responsible Use

This Actor returns article data and derived summaries from hosted storage. You are responsible for using the data in line with applicable laws, publisher terms, and your own compliance requirements.