An idea is only the first step
A view about a market still needs work before it becomes a strategy. Someone has to find the relevant markets, define conditions and check how the idea would have behaved in the past.
Then someone has to keep watching for those conditions.
AnyAlpha turns that sequence into an agentic trading platform. Users bring a thesis; the agent helps research it, build a strategy and backtest it before live execution. [1][2]
The company is starting with prediction markets, where relevant information can be scattered across news, data and social conversations. [2]
Research, test, then run
The process starts as a conversation. A user explains an idea and refines it with AnyAlpha into conditions the agent can follow. [2]
AnyAlpha identifies related markets and builds the strategy around the thesis. Historical backtesting lets the user inspect how those conditions would have behaved before choosing to go live. [1][2]
For a live strategy, the user connects an account. The agent monitors markets around the clock and places trades when the configured conditions are met. [1]
The iOS app also supports research, strategy discovery and paper trading for supported strategies. These give users ways to examine and refine an approach before committing capital. [3]
A user’s question about attention and price
One user brought AnyAlpha a thesis about sports prediction markets: heavy social attention might leave some NFL teams overpriced. [2]
The agent follows attention around teams on X, Reddit and TikTok. It compares market prices on Kalshi and Polymarket with estimated win probabilities, then defines conditions for a trade. [2]
The strategy is tested across historical seasons before the agent runs it live. Once active, the agent continues checking sentiment and prices against the rules. [2]
AnyAlpha keeps the original hypothesis, its signals and the execution conditions together as the user refines the strategy. Historical testing remains a separate step before live execution. [2]
Built by people who built trading systems
Shirley Peng and Jerry Miao founded AnyAlpha in 2026 and joined YC’s Fall 2026 batch. The two met at a poker tournament at Microsoft and later built apps together. [2]
Peng studied computer science at Waterloo and worked on trading infrastructure at Coinbase and a trading terminal at LSEG. Miao studied computer science at Berkeley and spent years building data, models, backtests and execution systems as a solo quant. [2]
They are applying that background to the gap between having an idea and operating a strategy. AnyAlpha combines the research and trading infrastructure that otherwise has to be assembled separately. [2]
A platform for creating and running agents
AnyAlpha’s starting point is prediction markets. Its stated direction is broader: agents that can be created, tested, run and eventually combined across more asset classes. [2]
The iOS app organizes research and strategy discovery around the user’s thesis. Paper trading is available for supported strategies alongside historical backtesting. [1][3]
Research does not end when an agent is created. The user can refine the original idea and examine the strategy, while the agent handles monitoring and execution under its conditions. [1][2]
Where AnyAlpha ranks
A ranked comparison is not yet available for AnyAlpha.
The bigger picture
AnyAlpha is built around a familiar gap: an interesting market idea rarely arrives with the data, rules and monitoring needed to act on it.
Its approach gives those steps one home, from a conversation about the thesis to a historical backtest and a running agent. [1][2]
The product’s ambition is to take the idea all the way from thesis to trade.
Written by SOTA2 from AnyAlpha’s public materials. Customer experiences and product plans are reported by AnyAlpha. Ranking analysis and comparison data are from SOTA2.