Blanket technical documentation
How Blanket turns a business outcome into one explainable market relationship, or an honest gap.
Blanket helps a business owner describe an outside event that could hurt the business, then searches current event markets for one useful relationship. The product keeps owner facts, market facts and deterministic dollar arithmetic separate.
Choose a path
| Reader | Start here | Then read |
|---|---|---|
| Business or product reviewer | Product scope | Analysis lifecycle and worked examples |
| Engineer | System overview | Market generations and public data boundary |
| Market designer | Demand overview | Evidence classes and publication lifecycle |
| Partner integrator | Partner API | Snapshot and sync and errors |
System contract
Blanket follows four durable rules:
- Owner facts and analyst suggestions remain distinct.
- The reasoning model can choose only from current candidates supplied by the application.
- Application code owns contract identity, market state, prices, quantities and arithmetic.
- Public outputs are rebuilt through explicit schemas and allowlists.
Product boundary
Blanket gives advisory analysis. It does not execute trades, hold money, sell insurance or promise that a market position will perfectly offset a business loss. The owner inspects and places any trade directly on Kalshi.
A weak relationship is not promoted to a position to fill the screen. The valid outcomes are a funded position candidate, an informative signal, an honest gap, a refusal, or a bounded error.