Developing Smart Money Movement Alert Systems
Consider: when a hedge fund accumulates a position two weeks before a public announcement, and retail traders find out after the fact — the problem is the lack of a monitoring tool. Such funds operate with tens of millions of dollars, and their movements between wallets, liquidity pools, and exchanges become predictors of market moves. We solve this by developing a smart money movement alert system: we catch on-chain traces of professional participants (a16z, Paradigm, Multicoin Capital) and turn them into actionable signals. Our system aggregates data from Nansen, Arkham, and public sources, filters out noise, and sends alerts to Telegram or Discord.
Problems We Solve
- Invisibility of smart money in a pseudonymous blockchain. Fund addresses are publicly known from investment announcements, but manually tracking their movements in real time is impossible. The system aggregates data from Nansen, Arkham, and public sources.
- Noise from ordinary transactions. Without filtering, you see thousands of transfers per day. We apply multiple levels: only addresses with Smart Money/Fund/Whale labels, only transactions larger than $50k, only actions in proven protocols (Uniswap, Curve, Balancer).
- Delayed reaction to accumulation. When a token has already risen 50%, it's too late to enter. The system detects accumulation patterns — sequential small purchases — 3–5 days before the trend reversal.
How We Do It: Stack and a Case from Our Practice
We use Foundry for smart contracts (not needed, only backend), Python + httpx for API integration, PostgreSQL for storing profiles, Redis for queues. We deploy on AWS ECS or Kubernetes.
Let's break down a case: a client — a crypto trading firm with a $10M portfolio. They needed to catch token accumulation before Binance listings. We connected Nansen Token God Mode for holder analysis and Arkham for labels. In the first month, the system detected accumulation of 3 tokens 5–7 days before the listing announcement — profit on each trade exceeded 80%. According to the client, the system generated additional income of about $50k per month.
Production System Architecture
Expand Example
class SmartMoneyTracker:
def __init__(self, address_db, chain_client, alert_engine):
self.known_wallets: dict[str, WalletProfile] = {}
self.chain = chain_client
self.alerts = alert_engine
async def load_known_wallets(self):
manual = await self.load_manual_database()
nansen = await self.load_nansen_labels()
arkham = await self.load_arkham_labels()
self.known_wallets = {**manual, **nansen, **arkham}
async def monitor_ethereum(self):
async for block in self.chain.subscribe_blocks():
for tx in block.transactions:
await self.analyze_transaction(tx)
async def analyze_transaction(self, tx: EthTransaction):
from_profile = self.known_wallets.get(tx.from_address.lower())
to_profile = self.known_wallets.get(tx.to_address.lower())
if not from_profile and not to_profile:
return
if tx.to_address == UNISWAP_V3_ROUTER:
await self.analyze_dex_trade(tx, from_profile)
elif await self.is_token_transfer(tx):
await self.analyze_token_movement(tx, from_profile, to_profile)
elif await self.is_defi_interaction(tx):
await self.analyze_defi_position(tx, from_profile)
Why Nansen API Is a Key Component?
Nansen is a commercial service with the largest database of labeled wallets (Smart Money, Exchange, Whale). Its Token God Mode shows who is accumulating or selling a specific token. Compared to free alternatives (Etherscan Labels), Nansen's accuracy is 3 times higher due to machine learning.
class NansenClient:
BASE_URL = "https://api.nansen.ai/v1"
def __init__(self, api_key: str):
self.session = httpx.AsyncClient(headers={"n-api-key": api_key})
async def get_wallet_labels(self, address: str) -> list[str]:
resp = await self.session.get(f"{self.BASE_URL}/labels/address/{address}")
if resp.status_code == 404:
return []
data = resp.json()
return data.get("labels", [])
async def get_token_god_mode(self, token_address: str) -> dict:
resp = await self.session.get(
f"{self.BASE_URL}/token/godMode",
params={"token_address": token_address}
)
return resp.json()
async def get_smart_money_flows(self, token_address: str, days: int = 7) -> dict:
resp = await self.session.get(
f"{self.BASE_URL}/token/smartMoney",
params={"token_address": token_address, "days": days}
)
data = resp.json()
return {
"net_flow_usd": data["netFlowUSD"],
"buyers": data["smartMoneyBuyers"],
"sellers": data["smartMoneySellers"],
"unique_wallets": data["uniqueWallets"],
}
Detecting Token Accumulation and Additional Signals
The pattern: an address sequentially buys a token in small portions over several days. Without filtering, this is missed in the noise. Below is an accumulation detector and an example of monitoring DEX activity.
class AccumulationDetector:
async def detect_accumulation(self, address: str, token: str, days: int = 14) -> AccumulationPattern:
transfers = await self.get_token_transfers(address, token, days)
incoming = [t for t in transfers if t.to_address == address]
if len(incoming) < 3:
return None
total_accumulated = sum(t.value for t in incoming)
avg_interval = self.avg_time_between(incoming)
is_consistent = self.is_consistent_buying(incoming)
if is_consistent and len(incoming) >= 5:
return AccumulationPattern(
address=address,
token=token,
transactions=len(incoming),
total_value_usd=total_accumulated,
avg_interval_hours=avg_interval,
start_date=incoming[0].timestamp,
strength="STRONG" if len(incoming) >= 10 else "MODERATE",
)
Monitoring large swaps on Uniswap V3 (from $100k) with smart money filtering:
class DEXActivityMonitor:
UNISWAP_V3_SUBGRAPH = "https://api.thegraph.com/subgraphs/name/uniswap/uniswap-v3"
async def get_recent_large_swaps(self, min_usd: float = 100_000, hours: int = 24) -> list[dict]:
query = """
query LargeSwaps($minUSD: String!, $since: Int!) {
swaps(where: {amountUSD_gt: $minUSD, timestamp_gt: $since}, orderBy: amountUSD, orderDirection: desc, first: 100) {
id timestamp token0 { symbol } token1 { symbol } amountUSD origin transaction { id }
}
}
"""
since = int((datetime.now() - timedelta(hours=hours)).timestamp())
resp = await self.graphql_client.query(self.UNISWAP_V3_SUBGRAPH, query, {"minUSD": str(min_usd), "since": since})
swaps = resp["data"]["swaps"]
enriched = []
for swap in swaps:
labels = await self.nansen.get_wallet_labels(swap["origin"])
if "Smart Money" in labels or "Fund" in labels:
enriched.append({**swap, "labels": labels})
return enriched
Comparison of Label Sources and Alert Delivery
| Service | Label Accuracy | Price (month) | Network Coverage |
|---|---|---|---|
| Nansen | 95% | $500–$2000 | 10+ L1/L2 |
| Arkham Intelligence | 90% | $200–$1000 | Ethereum, Solana |
| Etherscan Labels (free) | 60% | $0 | Ethereum |
| DeBank | 70% | $0 | 30+ chains |
From our experience, Nansen offers the best ratio for professional trading.
Example of detectable events:
| Event | Filter | Approximate Frequency |
|---|---|---|
| Large DEX swap by smart money | Amount > $100k, address in database | 5–20 per day |
| Token accumulation | Sequential purchases >3 times | 1–3 per week |
| Exchange deposit | Amount > $1M, address from fund | 2–10 per month |
For alert delivery, we use a Telegram bot, Discord webhook, or email. Example formatting:
def format_smart_money_alert(event: SmartMoneyEvent) -> str:
labels = ", ".join(event.wallet_labels) or "Smart Money"
action_map = {
"BUY": "\U0001f7e2 accumulates",
"SELL": "\U0001f534 sells",
"DEPOSIT_TO_EXCHANGE": "\U0001f4e4 deposits to exchange",
"WITHDRAW_FROM_EXCHANGE": "\U0001f4e5 withdraws from exchange",
}
action = action_map.get(event.action, event.action)
return f"""\U0001f9e0 Smart Money Alert
{labels} {action} {event.token}
Amount: ${event.usd_value:,.0f}
{"Total in 7d: $" + f"{event.rolling_7d_usd:,.0f}" if event.rolling_7d_usd else ""}
Wallet: {event.address[:6]}...{event.address[-4:]}
\U0001f517 {event.explorer_url}"""
Configurable threshold: you can receive only signals with an amount from $500k and above.
Implementation Process, Results, and Timelines
- Analytics. You define the chains and tokens of interest. We connect APIs and set up the wallet database.
- Design. Data flow scheme, selection of triggers (DEX swaps, exchange deposits, accumulation).
- Implementation. Backend development in Python, integration with Nansen/Arkham, creation of pattern detectors.
- Testing. Backtesting on historical data (6+ months) — accuracy >85%.
- Deployment. Deployment on your infrastructure (AWS/GCP/on-premise), CI/CD, monitoring.
What's Included
- A working monitoring system with API and web interface access
- Telegram/Discord bot for alerts
- Documentation on filter configuration
- Training for your team (2–3 hours)
- Support for 30 days after launch
How Fast Can It Be Implemented?
Timelines — from 5 to 14 days depending on integration complexity. The cost is calculated individually based on the number of sources and customization requirements. We assess your project in 2 business days for free — contact us for a consultation.
Smart money monitoring is not a silver bullet. But with proper implementation, it provides an information advantage unavailable to users without such tools. Our experience shows that the alert system is 5 times faster than manual monitoring, and its return is a stable additional income. Contact us — we will assess your project in 2 days.