Political Event Pulse (Realtime News -> Events + De-dup + AI-ready JSON)

1) Powerful Hook

Political monitoring breaks down when your pipeline is only a stream of article links.
Feeds are noisy, headlines repeat the same story, and raw outputs are hard to automate.
Political Event Pulse converts fragmented political news into structured, evolving events with source citations.
Instead of tracking articles, your team tracks what actually changed.

2) What Makes This Different From RSS / Google News

This Actor is not a discovery feed. It is a transformation and decision-support layer for production workflows.

Feature	RSS / Google News	Political Event Pulse
Article deduplication	No	Yes
Event clustering	No	Yes
Change detection between runs	No	Yes
Structured event JSON	No	Yes
AI-ready schema	No	Yes

RSS and Google News are optimized for reading. Political Event Pulse is optimized for monitoring systems, automation pipelines, and machine consumption.

3) Who This Is For (Clear Buyer Personas)

• Risk intelligence teams: Monitor geopolitical developments without manually triaging overlapping coverage.
• OSINT researchers: Build traceable event timelines from citation-backed records.
• AI builders (RAG pipelines): Ingest compact event objects instead of noisy article streams.
• Media monitoring startups: Deliver event-level products faster without building custom clustering infrastructure.
• Policy analysts: Track relevant developments with neutral, source-based summaries.
• Newsletter operators: Generate briefing candidates from event changes, not raw headline volume.

4) How It Works (Visual Flow)

Sources (RSS / Google News)
        ↓
Normalization
        ↓
Clustering
        ↓
Event Object
        ↓
Change Detection
        ↓
Dataset / Webhook / Telegram

Pipeline behavior

• Sources: RSS feeds, Google News RSS queries, optional start URL crawling.
• Normalization: URL cleanup, metadata shaping, optional content fetch.
• Clustering: Heuristic grouping (tokenJaccard or simhash) to collapse duplicate reporting.
• Change detection: Uses persisted fingerprints in KV store key STATE.
• Delivery: Emits to Apify Dataset and optionally sends webhook / Telegram notifications.

5) Event Object Design (Why It’s AI-Ready)

Each dataset item is one EventRecord, designed for both automation and analyst review.

• Stable eventId: Deterministic event hash supports incremental processing and joins.
• Compact neutral summary: Keeps context short and avoids persuasion-oriented language.
• Structured entities: Heuristic extraction of persons, orgs, and locations.
• Source citations: sourceUrls and sources[] preserve traceability.
• trendingScore: Lightweight prioritization based on recency and source characteristics.
• Automation-ready JSON: Works directly with LLM ingestion, RAG indexing, alert routing, and dashboards.

Dataset example

{
  "eventId": "f3f9f7b47f6e8de81385",
  "title": "Regional ministers discuss ceasefire framework",
  "summary": "Multiple reports describe renewed talks and proposed ceasefire terms.",
  "firstSeenAt": "2026-03-03T10:20:00.000Z",
  "lastUpdatedAt": "2026-03-03T12:05:00.000Z",
  "primarySource": "Reuters | example.com",
  "sourceCount": 4,
  "sourceUrls": ["https://example.com/a", "https://example.com/b"],
  "sources": [
    {
      "url": "https://example.com/a",
      "title": "Regional ministers discuss ceasefire framework",
      "publishedAt": "2026-03-03T12:00:00.000Z",
      "publisher": "Reuters",
      "language": "en"
    }
  ],
  "keywords": ["ceasefire", "minister", "talks"],
  "entities": {
    "persons": [],
    "orgs": ["Foreign Ministry"],
    "locations": ["tehran"]
  },
  "trendingScore": 8.12,
  "change": {
    "isNew": true,
    "isUpdated": false
  }
}

6) Real-World Use Cases (Concrete Examples)

• Monitoring sanctions: Group legal, diplomatic, and market-reaction coverage into single sanction-related events.
• Election tracking: Follow campaign, polling, and regulatory developments as evolving event objects.
• Conflict escalation alerts: Trigger incident workflows only when event state is new or materially updated.
• Parliamentary activity tracking: Aggregate legislative and executive actions for policy briefing pipelines.

7) Performance & Cost Positioning

Political Event Pulse is designed for operational use where reliability and cost control both matter.

• Cost-efficient default: Heuristic clustering and summarization run without external LLM cost.
• No mandatory AI dependency: Core output quality does not require OpenAI/Anthropic.
• Optional AI mode: You can enable LLM summarization when needed.
• Configurable volume: Use maxItemsPerRun, timeWindowHours, and source filters to manage runtime and spend.

Quick start input

If explicit sources are empty, keywords can be used as fallback Google News queries.

{
  "keywords": ["iran", "sanctions", "parliament"],
  "maxItemsPerRun": 100,
  "timeWindowHours": 48,
  "clusteringMethod": "tokenJaccard",
  "clusteringThreshold": 0.5,
  "changeDetectionEnabled": true,
  "emitMode": "newAndUpdated"
}

Webhook example

{
  "webhookUrl": "https://example.com/hooks/political-events",
  "webhookAuthType": "bearer",
  "webhookAuthToken": "YOUR_TOKEN",
  "webhookHeadersJson": "{\"X-Env\":\"prod\"}"
}

Webhook auth modes supported:

• none
• bearer
• basic
• apiKey

8) Safety & Compliance

• No fabricated content: The Actor transforms and summarizes source material only.
• Source-grounded output: Every event includes citations via sourceUrls and sources.
• Neutral presentation: Intended for monitoring and analysis, not political persuasion.
• Traceability first: Timestamps and source metadata are preserved for auditability.
• Redistribution responsibility: Users are responsible for applicable legal/commercial rules on stored or distributed content.

---

Political Event Pulse is built for teams that need structured political signal, not another feed reader.
If your workflow depends on event-level changes, machine-readable output, and source provenance, this Actor provides that layer out of the box.