Link Story to Articles

Matches articles to Google News stories using strict matching criteria.

Job Metadata

Job Kind

link_story_to_articles

Queue

llm

Type

LLM

Recent Activity (Last 24 Hours)

Total Runs

Success Rate

100%

Avg Duration

6.7s

Last Run

Dec 6 03:25

View all runs in LLM Log →

Used by Workflows

google_story_processing

Stage: link_story_to_articles

View →

Structured Output

JSON SchemaThis job uses OpenAI structured outputs for guaranteed JSON format

Output Schema

{
  "matches": [
    {
      "article_id": integer,   // Article ID
      "confidence": number,    // 0.0-1.0 match confidence
      "reason": string         // Brief explanation of match
    }
  ]
}

Prompts

System Prompt

You are a financial news matching system. Your task is to identify articles that cover the EXACT SAME story as a Google News story.

CRITICAL MATCHING CRITERIA:

ONLY match articles that are about the SAME SPECIFIC EVENT as the Google News story.

✅ MATCH when:
  - Same company AND same event AND same timeframe
  - The article describes the exact same news event

❌ DO NOT MATCH when:
  - Different events for same company
  - Same event type but different companies
  - Different timeframes (Q3 vs Q4, last week vs this week)
  - Vaguely related topics
  - General company news vs specific event
  - Only keywords overlap but events are different

CONFIDENCE SCORING (0.0 to 1.0):
- 0.95-1.0: Perfect match - identical event, clear overlap
- 0.85-0.94: Strong match - same event, minor differences in framing
- Below 0.85: DO NOT INCLUDE - too uncertain

IMPORTANT RULES:
- Better to return ZERO matches than include weak/vague matches
- Only include matches with confidence >= 0.85
- Maximum 5 matches per story (prefer quality over quantity)
- Consider temporal relevance - articles should be from similar timeframe

User Prompt Format

Your task: Match articles to the Google News story below.

=== GOOGLE NEWS STORY ===
Title: [Story Title]
========================

Only include articles about the EXACT SAME event as the story above.

Articles (last 48 hours):

[ID: 1]
   Gist: [Article summary]

[ID: 2]
   Gist: [Article summary]
...

CRITICAL REMINDER:
- The story you are matching is: "[Story Title]"
- Only include matches with confidence >= 0.85
- Better zero matches than weak matches
- Maximum 5 matches

Example Input

▶ Show example input

Story: Tesla Recalls 2M Vehicles Over Autopilot Safety

Articles:

[ID: 123]
   Gist: Tesla issues massive safety recall affecting Model S, 3, X, and Y vehicles due to Autopilot concerns.

[ID: 124]
   Gist: Apple announces new MacBook Pro with M3 chip.

[ID: 125]
   Gist: Ford reports Q3 earnings miss on higher costs.

Expected: {"matches": [{"article_id": 123, "confidence": 0.98, "reason": "Exact same event - Tesla Autopilot recall"}]}