FollowSM Official SDK (Python & TypeScript)

Official client libraries for the FollowSM Market Intelligence Engine — tracking Smart Money flows, Binance orderbook toxicity (VPIN), and Polymarket binary prediction confluence in real time.

🚀 Key Features

• Binance Microstructure: Tick-level VPIN, 1% orderbook toxicity bands, L1 imbalances, and volume Z-scores.
• Polymarket Confluence: Binary event odds, 15m probability deltas ($\Delta\text{Prob}$), and Smart Money whale sweeps.
• Cross-Market Divergence: Automated HFT risk recommendations ('NONE', 'WIDEN_SPREAD_1_5X', 'WIDEN_SPREAD_2X', 'HALT_MAKER_QUOTES')

Install

pip install followsm-sdk

Quickstart

from followsm_sdk import FollowSMClient

client = FollowSMClient(api_key="fsm_live_...")  # or omit for the free, IP-rate-limited tier
snapshot = client.get_toxicity_snapshot("BTCUSDT")
print(snapshot.vpin, snapshot.is_toxic_alert)

toxic_pairs = client.get_toxic_pairs()  # also works without an api_key

SymbolToxicityMetrics response

get_toxicity_snapshot() and get_toxic_pairs() both return SymbolToxicityMetrics (pydantic models), matching the raw JSON the backend returns:

{
  "symbol": "BTCUSDT",
  "timestamp": 1758412800.0,
  "price": 62150.5,
  "vpin": 0.72,
  "ob_toxicity_1pct": 2.35,
  "ob_imbalance_l1": 0.61,
  "depth_bands": {
    "0.5%": { "bid_notional": 184320.0, "ask_notional": 96410.0, "imbalance_ratio": 0.657 },
    "1.0%": { "bid_notional": 312500.0, "ask_notional": 210800.0, "imbalance_ratio": 0.597 },
    "2.0%": { "bid_notional": 590100.0, "ask_notional": 470200.0, "imbalance_ratio": 0.556 }
  },
  "volume_z_score": 3.1,
  "natr_15m": 0.84,
  "taker_buy_ratio": 0.58,
  "is_toxic_alert": true
}

Live streaming (Enterprise only)

stream_toxicity() connects to /ws/v1/toxicity, which requires an API key on an active Enterprise subscription — Developer API and free/unauthenticated keys are rejected with close code 4003.

import asyncio
from followsm_sdk import FollowSMClient

async def main() -> None:
    client = FollowSMClient(api_key="fsm_live_...")
    async for snapshot in client.stream_toxicity():
        print(snapshot.symbol, snapshot.vpin, snapshot.is_toxic_alert)

asyncio.run(main())

Cross-Venue Confluence (Binance × Polymarket)

get_confluence_snapshot(), get_confluence_snapshots() and stream_confluence() enrich Binance microstructure (VPIN, depth imbalance) with live Polymarket event flow (implied probability, CLOB order-flow imbalance, smart-money whale sweeps) and a composite HFT risk recommendation. They return ConfluenceSnapshot (pydantic models), matching this JSON:

{
  "symbol": "BTCUSDT",
  "timestamp_ms": 1790212800000,
  "binance_microstructure": {
    "price": 68420.50,
    "vpin": 0.78,
    "ob_toxicity_1pct": 2.14,
    "ob_imbalance_l1": 0.62,
    "depth_bands": {
      "0.5%": { "bid_notional": 450000, "ask_notional": 1200000, "imbalance_ratio": 2.66 },
      "1.0%": { "bid_notional": 1200000, "ask_notional": 2800000, "imbalance_ratio": 2.33 }
    },
    "volume_z_score": 3.1,
    "natr_15m": 0.87,
    "taker_buy_ratio": 0.29,
    "price_delta_15m_pct": 0.004
  },
  "polymarket_confluence": {
    "active_events": [
      {
        "market_slug": "will-btc-hit-70k-in-september",
        "question": "Will Bitcoin hit $70k in September?",
        "condition_id": "0x...",
        "yes_token_id": "12345...",
        "direction": "bullish_if_yes",
        "direction_confidence": 0.91,
        "implied_probability": 0.82,
        "prob_delta_15m": 0.09,
        "clob_order_flow_imbalance": 0.74,
        "smart_money_whale_sweeps_1h_usdt": 185000
      }
    ],
    "macro_event_risk_score": 0.85
  },
  "composite_signals": {
    "is_toxic_alert": true,
    "cross_market_divergence_flag": false,
    "recommended_action": "WIDEN_SPREAD_2X",
    "direction_ambiguous": false,
  }
}

snapshot = client.get_confluence_snapshot("BTCUSDT")
print(snapshot.composite_signals.recommended_action)

snapshots = client.get_confluence_snapshots(toxic_only=True)

Streaming (stream_confluence(), Enterprise only) connects to /ws/v1/confluence:

async for snapshot in client.stream_confluence():
    print(snapshot.symbol, snapshot.composite_signals.recommended_action)

Custom risk thresholds (RiskConfig)

The backend's recommended_action uses fixed, conservative thresholds. Quant clients can re-derive the risk ladder from the exposed binance_microstructure with their own thresholds via RiskConfig and evaluate_risk_action / client.evaluate_risk():

from followsm_sdk import FollowSMClient, RiskConfig

custom_config = RiskConfig(
    vpin_widen_threshold=0.65,      # Custom VPIN trigger for WIDEN_SPREAD_2X
    vpin_halt_threshold=0.85,       # Custom VPIN trigger for HALT_MAKER_QUOTES
    min_semantic_confidence=0.70,   # Stricter than the backend's 0.65 safety gate
)

client = FollowSMClient(api_key="YOUR_KEY", risk_config=custom_config)

snapshot = client.get_confluence_snapshot("BTCUSDT")
print(client.evaluate_risk(snapshot))  # re-evaluated with your own thresholds

evaluate_risk_action(snapshot, config) is also importable standalone for stateless/batch use. Like the backend, it never returns HALT_MAKER_QUOTES when the divergence rests on an event whose direction_confidence is below min_semantic_confidence.

Free vs Developer API

On HTTP 429, the SDK raises RateLimitExceededException with an upgrade prompt pointing to https://follow-sm.com/pricing.