Agentic research
Scrape prediction market data
Use a Notte Agent to research a prediction market and extract odds, prices, volume, and market context.
Run this template
Clone just this template, configure Notte, and start the run.
Before running
- Have
NOTTE_API_KEYready. Generate an API key.
Need help? Join the Notte Slack.
Research loop
- Uses a Notte Agent because the search and result-selection flow is open-ended and UI-dependent.
- Returns structured data with a Pydantic response model.
- Uses inline uv script metadata, so no template-specific pyproject.toml is required.
- Creates a Notte browser session.
Use it for
- Market research.
- Trading analysis.
- Prediction-market data enrichment.
Research caveats
- Missing credentials: verify .env contains NOTTE_API_KEY.
- Agent runs can take longer and cost more than deterministic scripts.
- Search queries may not map to active or visible markets.
- Kalshi may vary UI or gate content by region.