Agent-Generated Prediction Markets for Live-streaming

Whitepaper - April 2025


1. Executive Summary

whim.bet is a protocol for real-time, agent-curated prediction markets embedded directly into livestreams.

As livestreaming becomes the dominant medium for digital attention - whether in gaming, news, sports, or politics, viewers crave deeper forms of interaction. Today, this interactivity is shallow: limited to chat reactions, emojis, and algorithmic upvotes. Meanwhile, attention-rich moments - high-stakes plays, surprise announcements, narrative pivots - go economically unpriced.

whim.bet transforms these moments into dynamic markets.

Using a network of AI agents, the protocol continuously detects potential inflection points inside livestreams and autonomously generates prediction markets tailored to each moment. These markets are brief, contextual, and probabilistic—reflecting not the distant future, but the now-unfolding.

Rather than replacing entertainment, whim.bet reframes it as a field of ambient speculation.

The result is a new layer of interaction between viewer and stream:

The protocol runs atop agentic infrastructure and integrates seamlessly with no-code tooling such as auto.fun, making it deployable, extensible, and composable across a variety of use cases—from Twitch to town halls to tokenized esports.


2. Context & Rationale

2.1 The Cultural Shift

The livestream is the arena of the modern internet.

What television was in the 20th century, streaming has become today: a shared window into reality. But unlike television, streams are participatory, fast, and algorithmically fluid. They are not fixed experiences, but reactive ones.

And yet, interaction remains trivial: