TraderSpy Smart Money

TraderSpy Smart Money turns tracked account positions and ranking snapshots into evidence about trader behavior. TraderSpy covers Binance, Hyperliquid, Bybit, and OKX, with tools for an elite cross-exchange leaderboard, exchange-specific rankings, open or closed positions, individual profiles, position history, market statistics, and the current exchange list. It treats those observations as research rather than instructions to mirror another account.

Individual lookups always pair the trader ID with the exchange source from the originating row. Defaults can otherwise send an address or ID to the wrong exchange and return nothing. Symbol formats also differ: some rows use bare coins, others use USDT pairs, and tokenized stock symbols require fetching without a symbol filter and selecting them from the results. Symbols are read before aggregation rather than assumed equivalent.

Position rows expose side, coin size, entry, mark, leverage, timestamps, open state, and the last change event. Open rows carry unrealized profit or loss; closed rows carry realized results. Exposure summaries use approximate notional from size multiplied by mark price, not row counts. A coin-specific whale view totals notional by side, identifies the largest accounts with entry and leverage, and states the freshness of the newest observations. Closed rows can add recent exit context when open observations are sparse.

Use it for questions about whales, professional positioning, exchange rankings, or one trader's record. A coin-specific request can be:

What are whales doing in SOL?

Whole-market funding, open interest, and exchange-wide long-short ratios belong to the technical-analysis workflow. Automated signal research belongs to trading-signals. The connector sees only tracked accounts on its covered exchanges, so a visible long may be one leg of an unseen hedge. It cannot place trades or copy a trader.

Inputs include coin, exchange, ranking window, trader ID, position status, and pagination limits. Outputs are position tables or trader profiles with exchange, side, notional, entry, mark, leverage, result, and opening time. Rankings identify their time window and pair ROI with profit or loss and account size. Null win-rate fields stay unavailable rather than becoming inferred values.

The smart score includes realized performance, effective win rate, ROI edge, consistency, longevity, trade depth, recency, and staleness adjustments. Its breakdown, rationale, and raw metrics explain the ranking. A reliability multiplier below one flags short-history limitations. Profile research can calculate closed-trade win rate, typical holding time and leverage, concentration, worst trade, and recent-versus-all-time behavior.

Aggregate market statistics include traders who fell off leaderboards, so large aggregate losses do not become claims that the current leaders are losing. Past rankings and win rates describe their samples, not future winners.

get_elite_leaderboard returns the cross-exchange top 10. get_top_traders supports 24h, 3D, 7D and 30D windows with a limit up to 200; get_positions and get_trader_position_history cap a page at 50. get_trader_profile and history must receive both traderId and source. get_market_stats supports 4h, 8h, 24h and 7d, and get_exchanges identifies covered venues. A BTC symbol filter matches BTC and BTCUSDT exactly; xyz:AVGO requires an unfiltered fetch.

The daily 0-100 smart score weights realized PnL 30%, effective win rate 22%, ROI edge 18%, consistency 14%, longevity 10%, and trade depth 6%. Those six weights sum to 100%. Recency can add up to 9 and staleness subtract up to 12 after about three weeks idle. The elite run occurs at 00:10 Europe/Istanbul; lastRunAt identifies freshness. algorithmDetails supplies the actual weights, while totalTrackedTraders and tradersWithClosedTrades distinguish coverage from closed-trade evidence. A high score may coexist with negative effectiveRoi when realized PnL dominates. Venue badges can be null.

Binance position snapshots refresh every 5-10 minutes, Bybit every 10-15 minutes, Hyperliquid every 15 minutes plus live fills, and OKX every 10-60 minutes. Binance leaders opted into public visibility; Hyperliquid records are on-chain and winRate is null; OKX includes copy-trade leaders whose high win rates can accompany tiny per-trade ROI. Free position rows are delayed 15 minutes, while profiles, leaderboards and stats are not delayed.

A research profile can use two pages of 50 closed rows and state its sample size. Win rate counts positive-pnl rows divided by all rows; average hold uses closeTime minus openTime. Concentration measures share of rows or absolute pnl in the most-traded symbol. Compare minimum pnl with median wins, flag median leverage of at least 20x, and use the share of rows from the last 30 days to expose dormant records. These are observations and cannot execute or copy a trader.

The hosted endpoint is https://mcp.traderspy.app/mcp over Streamable HTTP. Connect it in the client and authorize with OAuth or a personal key from https://traderspy.app/mcp. A free account is sufficient. Daily tool limits are 300 calls on the free tier and 5,000 on premium. Without the server connection, this skill has no live market data.