Practical Prediction Markets MCP workflows for market discovery, liquidity monitoring, historical backfills, and exchange-level analytics.

Prediction Markets MCP Use Cases

This page shows practical ways to combine the Prediction Markets MCP tools into repeatable workflows.

1. Discover active markets on one exchange

Use this workflow when you want to find valid market_id values quickly and keep the first response compact.

Step 1: list exchanges

{
  "tool": "exchanges_list",
  "arguments": {}
}

Step 2: list active market IDs

{
  "tool": "markets_get_active",
  "arguments": {
    "exchange_id": "POLYMARKET",
    "limit": 100,
    "page": 1
  }
}

Step 3: switch to rich market objects if needed

{
  "tool": "markets_get_history",
  "arguments": {
    "exchange_id": "POLYMARKET",
    "limit": 25,
    "page": 1
  }
}

Why this works well

markets_get_active is the fastest way to discover valid market IDs.
markets_get_history adds title, description, status, and price context only when needed.

2. Monitor current liquidity for one market

Use this workflow when you want the freshest market state for dashboards, agents, or monitoring jobs.

Step 1: get current market activity

{
  "tool": "activity_get_current",
  "arguments": {
    "exchange_id": "POLYMARKET",
    "market_id": "1-MEGATON-METEOR-STRIKE-IN-2026_NO"
  }
}

Step 2: get the current order book

{
  "tool": "orderbook_get_current",
  "arguments": {
    "exchange_id": "POLYMARKET",
    "market_id": "1-MEGATON-METEOR-STRIKE-IN-2026_NO"
  }
}

What you get

recent trade and quote context from activity_get_current
current bids and asks from orderbook_get_current
enough information to compute spread, depth, and top-of-book imbalance

3. Backfill candles for one market

Use this when you want compact time-bucketed history instead of raw event streams.

Step 1: discover supported periods

{
  "tool": "ohlcv_list_periods",
  "arguments": {}
}

Step 2: fetch the latest candles

{
  "tool": "ohlcv_get_latest_market",
  "arguments": {
    "exchange_id": "POLYMARKET",
    "market_id": "1-MEGATON-METEOR-STRIKE-IN-2026_NO",
    "period_id": "1DAY",
    "limit": 10
  }
}

Step 3: request a bounded historical window

{
  "tool": "ohlcv_get_history_market",
  "arguments": {
    "exchange_id": "POLYMARKET",
    "market_id": "1-MEGATON-METEOR-STRIKE-IN-2026_NO",
    "period_id": "1DAY",
    "time_start": "2026-04-01T00:00:00Z",
    "time_end": "2026-04-07T00:00:00Z",
    "limit": 100
  }
}

Why this works well

candles are easier to chart and cache than raw trades or quotes
ohlcv_get_latest_market is good for refresh loops
ohlcv_get_history_market is better for deterministic backfills

4. Pull raw trade or quote history with bounded windows

Use this workflow when you need flat-file event history rather than summaries.

Example: trades history

{
  "tool": "activity_get_trades_history",
  "arguments": {
    "exchange_id": "POLYMARKET",
    "market_id": "1-MEGATON-METEOR-STRIKE-IN-2026_NO",
    "time_start": "2026-04-07T12:00:00Z",
    "time_end": "2026-04-07T13:00:00Z",
    "limit": 100
  }
}

Example: quotes history

{
  "tool": "activity_get_quotes_history",
  "arguments": {
    "exchange_id": "POLYMARKET",
    "market_id": "1-MEGATON-METEOR-STRIKE-IN-2026_NO",
    "time_start": "2026-04-07T12:00:00Z",
    "time_end": "2026-04-07T13:00:00Z",
    "limit": 100
  }
}

Typical response shape:

[
  {
    "entry_time": "2026-04-07T13:35:12.0000000Z",
    "recv_time": "2026-04-07T13:35:12.1230000Z",
    "ask": 0.949,
    "bid": 0.928,
    "ask_volume": 112.39,
    "bid_volume": 81.55
  }
]

Notes

these tools are best used with narrow windows
results are filtered by time_coinapi
start with a small limit, then widen only if needed

5. Reconstruct book changes from historical order updates

Use this when you need event-level order book evolution instead of a single snapshot.

{
  "tool": "orderbook_get_history",
  "arguments": {
    "exchange_id": "POLYMARKET",
    "market_id": "1-MEGATON-METEOR-STRIKE-IN-2026_NO",
    "time_start": "2026-04-07T12:00:00Z",
    "time_end": "2026-04-07T12:10:00Z",
    "limit": 100
  }
}

This is especially useful if:

you maintain your own order book reconstruction logic
you want to replay order adds, updates, and deletes
current snapshots are not enough for your analysis

6. Build exchange-wide short-window analytics

Use this when you want one OHLCV window across all markets on an exchange.

{
  "tool": "ohlcv_get_history_exchange",
  "arguments": {
    "exchange_id": "POLYMARKET",
    "period_id": "1MIN",
    "time_start": "2026-04-07T12:00:00Z",
    "time_end": "2026-04-07T13:00:00Z"
  }
}

Best fit

venue-level scans
cross-sectional ranking
exchange-wide dashboards for a short time window

Notes

market_id values include the outcome, so store them exactly as returned by discovery tools.
markets_get_active is usually the cheapest entry point for discovery, but markets_get_history is better when you need human-readable metadata.
Keep historical flat-file requests narrow to avoid unnecessarily large responses.
ohlcv_get_history_exchange is capped at a 24-hour range.

Service Status GitHub SDK

Use Cases

On this page