Introduction
Use the FinFeedAPI Prediction Markets MCP server to discover exchanges, inspect markets, query activity, read order books, and pull OHLCV timeseries through self-describing tools.
Prediction Markets API MCP
The FinFeedAPI Prediction Markets API is available through a hosted Model Context Protocol (MCP) server at https://api.prediction-markets.finfeedapi.com/mcp.
It exposes the API as self-describing tools, which makes it easy to use from MCP clients such as Cursor, Claude Desktop, or custom agent runtimes.
Use your regular FinFeedAPI API key in the X-APIKey header.
What is available through MCP?
The Prediction Markets MCP server currently exposes 14 tools grouped into four areas:
| Area | What you can do |
|---|---|
| Exchange and market discovery | List supported exchanges, inspect a single venue, enumerate active market IDs, or retrieve full market records. |
| Market activity | Read the latest trade and quote, fetch recent activity batches, or retrieve historical trades and quotes from flat files. |
| OHLCV timeseries | Inspect supported candle periods, fetch latest candles for a market, retrieve historical candles for a market, or request exchange-wide OHLCV windows. |
| Order books | Read the current order book snapshot or retrieve historical limit order updates from flat files. |
Typical discovery flow
Most integrations follow this order:
- Call
exchanges_listto discover supported venues such asPOLYMARKET,KALSHI,MYRIAD, orMANIFOLD. - Call
exchanges_getif you want venue metadata for one exchange. - Call
markets_get_activefor a compact list of activemarket_idvalues, ormarkets_get_historyfor fuller market objects with titles, descriptions, and prices. - Use
activity_get_currentororderbook_get_currentfor the latest market state. - Call
ohlcv_list_periods, then useohlcv_get_latest_marketorohlcv_get_history_marketfor candles. - Use
activity_get_trades_history,activity_get_quotes_history, ororderbook_get_historyonly when you need flat-file history and can keep the requested range narrow.
Live-tested notes
markets_get_activereturns only market IDs, not full market objects.ohlcv_get_latest_marketreturns rows in descending time order.ohlcv_get_history_marketandohlcv_get_history_exchangereturn rows in ascending time order.ohlcv_get_history_exchangerequires bothtime_startandtime_end, and the requested window cannot exceed 24 hours.- Historical flat-file tools support either a full-day
datequery or a boundedtime_start/time_endrange. - For low-activity markets, current activity calls may return sparse results or an explicit "no activity found" error.
orderbook_get_currentreturns a snapshot withbidsandasks, whileorderbook_get_historyreturns raw order updates rather than reconstructed snapshots.
Start here
Getting Started
Configure an MCP client, authenticate requests, and verify your Prediction Markets MCP connection.
Tool Reference
Full reference for all Prediction Markets MCP tools, parameters, workflows, and usage patterns.
Use Cases
Practical MCP workflows for market discovery, monitoring, historical backfills, and liquidity analysis.
