← docs

Prediction markets and strategy

Five tools cover the full intelligence loop: discover markets, assess liquidity, detect actionable signals, and surface composite strategies across all Syenite data sources. Prediction data comes from Polymarket; strategy search combines yield, carry, lending, prediction signals, and gas estimates.

prediction.trending

Top markets by volume — a quick read on what the crowd is pricing.

Parameters:

Response: event title, slug, total volume (USDC), liquidity, and per-outcome probabilities, best bid/ask, spread. Use this as a starting point to find interesting markets, then drill into specific outcomes with prediction.book.

prediction.search

Search markets by keyword — "Bitcoin", "Fed rate", "election", or any topic.

Parameters:

Same response shape as prediction.trending. Combine with signals to find tradeable opportunities in a specific domain.

prediction.book

Order book depth for a single outcome token. Use to assess execution cost, slippage, and available depth before trading.

Parameters:

Response: bids, asks, spread, mid-price, cumulative depth at various price levels.

prediction.signals

The detection layer. Scans Polymarket for actionable opportunities: wide spreads (market-making), extreme probabilities (mean-reversion), volume spikes (momentum), deep liquidity pools, and mispriced outcomes.

Parameters:

Each signal includes the market context, signal type, strength score, and a suggested action. Agents can use this for autonomous strategy discovery — scan for signals, filter by type, then execute via the Polymarket CLOB or hedge on-chain.

Example: "Find wide-spread signals above strength 60" → prediction.signals with types: ["wide_spread"], minStrength: 60. The agent identifies markets where market-making is profitable, checks depth via prediction.book, and executes.

find.strategy

The intelligence layer that ties all Syenite data sources together. Tell it what asset you have and your risk tolerance; it scans yield opportunities, carry trades, lending leverage, prediction signals, and gas costs to return ranked strategies with expected returns and execution steps.

Parameters:

Response: ranked strategies, each with category (yield, carry, leverage, prediction, arbitrage), expected return, risk level, confidence, and the specific Syenite tool calls to execute. The agent can chain from find.strategy directly into lending.supply, swap.quote, or any other execution tool.

Example: "I have 50k USDC on Ethereum, medium risk" → find.strategy with asset: "USDC", amount: 50000, riskTolerance: "medium", chain: "ethereum". Returns a ranked list: best yield vault, top carry trade, and any prediction signal plays — with exact next steps.

typical flow

1. prediction.trending or prediction.search to discover markets. 2. prediction.book on interesting outcomes for depth. 3. prediction.signals to scan for actionable opportunities. 4. find.strategy to get a holistic view combining prediction signals with yield, carry, and gas data. 5. Execute with lending.supply, swap.quote, or external trading tools.

related

Yield opportunities. Lending rates and risk. Lending execution. Position alerts and carry. Full tool list.