Build a News Sentiment Oracle on GenLayer — Tutorial

Prerequisites

• GenLayer Studio account at studio.genlayer.com
• Connected To Genlayer Studio Network
• Test GEN tokens (use the Faucet)

Step 1: Create the Contract

Open Studio → click "+" → name it news_sentiment_oracle.py

Paste this code:

# v0.1.0
# { "Depends": "py-genlayer:15qfivjvy80800rh998pcxmd2m8va1wq2qzqhz850n8ggcr4i9q0" }
from genlayer import *
import json


class NewsSentimentOracle(gl.Contract):
    """
    News Sentiment Oracle — Intelligent Contract

    Fetches live news about any topic and analyzes sentiment on-chain.
    AI validators independently fetch and evaluate news, then reach
    consensus on whether the sentiment is bullish, bearish, or neutral.

    Real-world use cases:
    - Trigger DeFi trades based on news sentiment
    - DAO governance — only execute proposal if community sentiment is positive
    - Prediction markets — resolve based on news coverage
    - Risk management — pause protocol if negative news detected
    """
    topic: str
    sentiment: str
    confidence: str
    headline: str
    last_updated: str

    def __init__(self):
        self.topic = ""
        self.sentiment = "unknown"
        self.confidence = ""
        self.headline = ""
        self.last_updated = ""

    @gl.public.write
    def analyze_sentiment(self, topic: str) -> None:
        """
        Fetch latest news about a topic and analyze sentiment.

        Args:
            topic: Any topic e.g. "bitcoin", "ethereum", "GenLayer", "Apple stock"
        """
        def fetch():
            url = "https://news.google.com/search?q=" + topic + "&hl=en"
            return gl.get_webpage(url, mode="text")

        result = gl.eq_principles.eq_principle_prompt_non_comparative(
            fetch,
            task="Read the latest news headlines about " + topic + ". Analyze the overall sentiment. Is the news mostly positive/bullish, negative/bearish, or neutral? Also extract the single most important headline. Respond with: sentiment|confidence|headline. Example: bullish|high|Bitcoin hits new all-time high as institutional demand surges",
            criteria="Answer must follow format: sentiment|confidence|headline where sentiment is bullish, bearish, or neutral, and confidence is high, medium, or low."
        )

        parts = str(result).strip().split("|")
        if len(parts) >= 3:
            self.sentiment = parts[0].strip().lower()
            self.confidence = parts[1].strip().lower()
            self.headline = parts[2].strip()
        else:
            self.sentiment = str(result).strip()
            self.confidence = "low"
            self.headline = "unable to extract"

        self.topic = topic
        self.last_updated = "latest"

    @gl.public.write
    def check_bullish(self, topic: str) -> None:
        """
        Simple bullish check — useful for triggering contract logic.
        Sets sentiment to 'true' if bullish, 'false' otherwise.

        Args:
            topic: Topic to check e.g. "bitcoin", "ethereum"
        """
        def fetch():
            url = "https://news.google.com/search?q=" + topic + "&hl=en"
            return gl.get_webpage(url, mode="text")

        result = gl.eq_principles.eq_principle_prompt_non_comparative(
            fetch,
            task="Based on the latest news headlines about " + topic + ", is the overall sentiment bullish/positive? Respond only: true or false",
            criteria="Answer must be exactly: true or false"
        )
        self.topic = topic
        self.sentiment = str(result).strip().lower()
        self.confidence = "high"
        self.headline = ""

    @gl.public.view
    def get_sentiment(self) -> str:
        return self.sentiment

    @gl.public.view
    def get_confidence(self) -> str:
        return self.confidence

    @gl.public.view
    def get_headline(self) -> str:
        return self.headline

    @gl.public.view
    def get_topic(self) -> str:
        return self.topic

Step 2: Deploy

Click Deploy — no parameters needed. Wait for Finalized status.

Save your Transaction Hash.

Step 3: Analyze News Sentiment

Call analyze_sentiment (Write method):

topic → "bitcoin"

Wait for Finalized (~30-60 seconds).

Then call get_sentiment (Read):

Result: "bullish" or "bearish" or "neutral"

Call get_headline (Read) to see the top news headline that influenced the verdict.

Step 4: Use the Binary Check

For contract logic that needs a simple true/false trigger, call check_bullish:

topic → "ethereum"

Then get_sentiment returns "true" or "false" — ready to use in conditional logic.

How It Works Under the Hood

1. You call analyze_sentiment("bitcoin")
2. Each GenLayer validator independently:
   → Fetches Google News for "bitcoin"
   → Sends content to their LLM
   → LLM evaluates sentiment
3. Validators compare results via Equivalence Principle
   → If majority agree → consensus reached
   → Result stored on-chain
4. You read get_sentiment() → "bullish"

No oracle. No trusted third party. Just AI validators reaching consensus.

Real-World Use Cases

if oracle.get_sentiment() == "bullish":
    execute_buy_order()

if oracle.get_sentiment() != "bearish":
    pass_proposal()

if oracle.get_sentiment() == "bearish" and oracle.get_confidence() == "high":
    pause_protocol()

Built as part of the GenLayer Incentivized Builder Program.