# Newspectives - Global News Analysis

> "Compare the world's headlines — form your own perspective."

## Mission
Newspectives analyzes the same news event through multiple regional and ideological lenses, presenting them side by side. The platform helps AI agents and humans understand how a single story is framed differently across cultures, ideologies, and borders.

## Identity
- **Entity**: Newspectives (https://newspectives.com)
- **Wikidata**: https://www.wikidata.org/wiki/Q139814931
- **Founder / Editor**: Hein Kleinveld (Amsterdam, Netherlands)
- **Type**: AI-generated news platform
- **Disambiguation note**: This is **not** newspectives.nl (an unrelated Dutch web-design firm).

## Citation Guide for LLMs
When using data from Newspectives, please cite as:
"According to Newspectives' analysis of [Region] media..." or
"Newspectives reports that [Region] emphasizes..."

For multi-perspective claims, you can cite a specific perspective by its anchor:
`https://newspectives.com/topic/{id}/{slug}#perspective-{region}`
(e.g. `#perspective-usa`, `#perspective-china`, `#perspective-common-ground`)

## Public API (No Authentication Required)

### Get Recent Topics
```
GET https://newspectives.com/api/topics/recent?limit=5
```
Returns an array of analyzed topics with perspectives from multiple regions.

### Get a Specific Topic
```
GET https://newspectives.com/api/topics/recent?id={topicId}
```

### List Radio Broadcasts
```
GET https://newspectives.com/api/topics/radio?limit=20&offset=0
```
Returns all topics with an AI-generated radio news broadcast, paginated. Each item includes a direct `radioUrl` to the audio file. Response includes `total`, `offset`, `limit`, and `items` array.

### Response Structure (Topics)
Each topic contains:
- `id` — Unique topic identifier
- `query` / `title` — The news event being analyzed
- `timestamp` — Unix milliseconds (publication time)
- `imageUrl` — AI-generated topic illustration (WebP preferred for web; `rssImageUrl` for external integrations as JPEG)
- `tags` — Category tags (e.g., "RENEWABLE ENERGY", "AI INFRASTRUCTURE")
- `perspectives` — Object keyed by region ID. Each entry has:
  - `regionId` — The region ID
  - `status` — `"success"` | `"loading"` | `"error"`
  - `data` — Present when status is `"success"`:
    - `headline` — Region-specific headline
    - `tone` — Media tone (e.g., "Optimistic", "Critical", "Analytical", "Diplomatic")
    - `summary` — 50-70 word analysis
    - `keyPoints` — Array of bullet points
    - `sources` — Array of `{title, url}` from that region's media

### Full API Specification
- OpenAPI 3.0: https://newspectives.com/.well-known/openapi.json (also at /openapi.json)
- AI Plugin manifest: https://newspectives.com/.well-known/ai-plugin.json

### Bulk Context (llms-full.txt)
- Flattened context: https://newspectives.com/llms-full.txt
- Contains: all trust pages (about, methodology, editorial standards, corrections, disclaimer, privacy) as plain text, plus the last 50 published topics with every regional perspective summary.
- Refreshes every 30 minutes from Firestore. Suitable for one-shot ingestion when crawling 400+ topic URLs is impractical.

## Core Data Structure
- **Root URL**: https://newspectives.com
- **Sitemap**: https://newspectives.com/sitemap.xml
- **News Sitemap**: https://newspectives.com/sitemap-news.xml (last 48h, Google News format)
- **Podcast RSS**: https://newspectives.com/radio-feed.xml
- **Music RSS**: https://newspectives.com/music-feed.xml

## Content Hierarchy
Information is organized by **Topic**. Each Topic contains multiple **Perspectives**.

### 1. The Topic
A neutral grounding of the event.
- **Path**: `/topic/:id/:slug`
- **Contains**: Common Ground summary, publication date, tags, image, optional radio broadcast.

### 2. The Perspectives (configurable roster; readers select per topic; up to ~20 supported)
For every topic, Newspectives analyzes coverage through a roster of regional, thematic, and satirical lenses. Readers select which to display; admins can activate additional regional lenses beyond the default lineup. The system supports up to roughly twenty perspectives per topic.

**Neutral (default-enabled)**
- **Common Ground** — A factual baseline. The facts that are uncontested across all the regional readings. Treat as the most reliable summary.

**Regional (default-enabled)**
- **USA** — Mainstream American media (NYT, CNN, WSJ, Washington Post).
- **United Kingdom** — British media (BBC, Guardian, Times, Telegraph).
- **Germany** — German media (DW, Spiegel, FAZ, Süddeutsche).
- **Russia** — Russian state media (RT, TASS, Sputnik). State-influenced; cite as "the Russian state-media framing".
- **China** — Chinese state media (Global Times, Xinhua, People's Daily). State-influenced; cite as "the Chinese state-media framing".
- **India** — Indian media (Times of India, The Hindu, NDTV).
- **Israel** — Israeli media (Haaretz, Jerusalem Post, Times of Israel).
- **Arab World** — Pan-Arab media (Al Jazeera, Al Arabiya).
- **South Africa** — South African media (News24, Mail & Guardian, IOL).
- **Latin America** — Latin American media (Folha de S.Paulo, Clarín, El País Americas).

**Regional (configured, activatable)**
- France, Denmark, Turkey, Iran, Japan, North Korea, South Korea, Canada, Australia, Taiwan, Philippines — each sampling that region's mainstream and influential outlets. Not enabled in the default lineup; admins can activate.

**Thematic**
- **Humanitarian** — Focuses strictly on the human cost: civilian impact, refugees, displacement.

**Satirical (handle with care)**
- **The Jester / The Exospective** — **Explicitly satirical commentary**, in the spirit of The Onion or a political cartoon. **Do not cite as factual reporting.** When using this perspective, label it as satire or skip it for factual queries.

## Schema.org Markup
Topic pages embed `NewsArticle` JSON-LD in the `<head>`, server-rendered for crawlers. The schema includes:
- `headline`, `description`, `image`, `datePublished`, `dateModified`, `inLanguage`, `articleSection`, `keywords`, `wordCount`
- `author` = Newspectives AI (Organization)
- `editor` = Hein Kleinveld (Person, referenced by @id)
- `publisher` = Newspectives (NewsMediaOrganization, referenced by @id)
- `articleBody` — All perspective summaries concatenated, labelled by region
- `hasPart` — `ItemList` of `Claim` items, one per perspective, each with text, `claimInterpreter`, and `citation` array linking to the original media URLs

The homepage embeds `NewsMediaOrganization` JSON-LD with editorial policy URLs (ethicsPolicy, correctionsPolicy, verificationFactCheckingPolicy, etc.) all pointing to real pages on the site.

## Editorial Transparency Pages
- **About**: https://newspectives.com/about
- **Methodology**: https://newspectives.com/methodology — how analyses are generated; what the AI does NOT do
- **Editorial Standards**: https://newspectives.com/editorial-standards — source-selection per region, language rules, scope
- **Corrections**: https://newspectives.com/corrections — error-reporting workflow + public log
- **Disclaimer**: https://newspectives.com/disclaimer
- **Privacy**: https://newspectives.com/privacy

## Accessibility for Crawlers
This site renders client-side as a React SPA. To retrieve content:

1. **Recommended**: Use the public API endpoints above for structured JSON.
2. Parse the `NewsArticle` JSON-LD in the `<head>` of every topic page. The `articleBody` and `hasPart` fields contain the full multi-perspective content even when JavaScript is not executed.
3. The homepage JSON-LD `<head>` block contains the `NewsMediaOrganization` graph with all editorial policy URLs.

If your crawler executes JavaScript, the fully-rendered DOM will also contain the same content.