mackinac-client

Python client for the Mackinac market-data API — live quotes, trades, funding rates, yield rates, and lending markets across 16 DeFi venues including Hyperliquid, dYdX V4, Uniswap V3/V4, Pendle, Spectra, GMX, Vertex, Ostium, Aave V3, Compound V3, Morpho Blue, and more.

pip install mackinac-client

---

Quick start

Synchronous (no asyncio required)

No credentials required for free-tier access (3 concurrent symbols):

from mackinac import Mackinac, QuoteMessage, PrintMessage

with Mackinac() as m:
    # What symbols are on HL?
    print(m.live_symbols("hl"))

    # Stream only quotes — types= filters the feed, no isinstance needed
    with m.subscribe("hl:ETH", types=QuoteMessage) as feed:
        for quote in feed:
            bid = quote.bids[0].price if quote.bids else None
            ask = quote.asks[0].price if quote.asks else None
            print(f"ETH  bid {bid}  ask {ask}")
            break  # just the first one

    # Stream only trades
    with m.subscribe("hl:ETH", types=PrintMessage) as feed:
        for trade in feed:
            side = "buy" if trade.side == 1 else "sell"
            print(f"ETH  {side}  {trade.price}  ×  {trade.size}")
            break

Async (for existing asyncio code)

import asyncio
from mackinac import AsyncClient, QuoteMessage

async def main():
    async with AsyncClient() as client:
        print(await client.live_symbols("hl"))

        async with client.subscribe("hl:ETH", types=QuoteMessage) as feed:
            async for quote in feed:
                print(f"ETH  bid {quote.bids[0].price}  ask {quote.asks[0].price}")
                break

asyncio.run(main())

---

Authentication

Both Mackinac (sync) and AsyncClient (async) share the same constructors:

API key (recommended for servers)

m      = Mackinac.from_api_key("mk_live_...")       # sync
client = AsyncClient.from_api_key("mk_live_...")    # async

API keys are long-lived and authenticate WebSocket connections only. Obtain them from the dashboard (requires tier api).

JWT (browser / short-lived sessions)

m      = Mackinac.from_jwt("eyJhbGciOi...")
client = AsyncClient.from_jwt("eyJhbGciOi...")

JWTs expire after 7 days. Refresh via client.refresh_token() / await client.refresh_token().

Wallet sign-in (EIP-191)

pip install 'mackinac-client[wallet]'

m      = Mackinac.from_wallet(private_key="0xdeadbeef...", address="0xYourWalletAddress")
client = await AsyncClient.from_wallet(
    private_key="0xdeadbeef...",
    address="0xYourWalletAddress",
)

---

Subscribing to live data

async with client.subscribe("hl:ETH", "uni:WETH/USDC", "rates:all") as feed:
    async for msg in feed:
        match msg.type:
            case "quote":
                print(msg.exchange, msg.symbol, msg.bids[0].price)
            case "print":
                print(msg.price, msg.size, msg.side)  # side: 0=sell 1=buy
            case "rate_market":
                print(msg.symbol, f"{msg.impliedApy * 100:.2f}% APY")
            case "funding":
                print(msg.symbol, f"{msg.ratePct:.4f}% ann.")

All messages are typed Pydantic v2 models. Use isinstance() or match on msg.type to branch per message type.

Subscribe key format

Venue	Key format	Example
Hyperliquid perp	hl:<SYMBOL>	hl:ETH
dYdX V4 perp	dydx:<BASE>-USD	dydx:ETH-USD
Uniswap V3 (Arbitrum)	uni:<BASE>/<QUOTE>	uni:WETH/USDC
Uniswap V4 (Arbitrum)	univ4:<BASE>/<QUOTE>	univ4:WETH/USDC
SushiSwap V3	sushi:<BASE>/<QUOTE>	sushi:WETH/USDC
PancakeSwap V3	pancake:<BASE>/<QUOTE>	pancake:WBTC/WETH
GMX perp	gmx:<SYMBOL>	gmx:BTC
Vertex perp	vertex:<SYMBOL>	vertex:SOL
Ostium (RWA)	ostium:<BASE>/<QUOTE>	ostium:XAU/USD
Aave V3 lending	aave:<ASSET> or aave:all	aave:USDC
Compound V3 lending	compound:<ASSET> or compound:all	compound:USDC
Morpho Blue lending	morpho:<ASSET> or morpho:all	morpho:USDC
Lending actions (all venues)	lending:actions	lending:actions
Pendle/Spectra (all)	rates:all	—
Pendle/Spectra (trades)	rates:swaps	—
Specific yield market	rates:<symbol-or-address>	rates:PT-weETH-25JUN2026
AMM consolidated NBBO	ammbook:<BASE>/<QUOTE>	ammbook:WETH/USDC (professional+)

Use mackinac.symbols helpers to construct keys safely:

from mackinac import symbols

symbols.amm_pair("WETH", "USDC")        # "WETH/USDC"
symbols.ostium_pair("XAU")              # "XAU/USD"
symbols.pendle_address("0xC62D...")     # "0xc62d..."  (lowercased)
symbols.rates_all()                     # "rates:all"

Unsubscribing mid-session

await feed.unsubscribe("hl:ETH")

Auto-reconnect

The library reconnects automatically on drops (1s → 2s → 4s → … → 30s cap) and re-subscribes all active symbols. The async for loop continues seamlessly — no user-side retry logic needed.

Raw websockets note

If you connect with the websockets library directly (rather than via AsyncClient), set max_size to at least 10 MB. The initial trade snapshot sent on subscribe can exceed the library's 1 MB default:

async with websockets.connect(url, max_size=10 * 1024 * 1024) as ws:
    ...

AsyncClient already sets this internally.

---

REST methods

Symbol discovery

# Which exchanges are live?
await client.markets()                        # dict[exchange, MarketStatus]

# All subscribable symbols on one venue
await client.live_symbols("hl")              # list[str]
await client.live_symbols("pendle")          # list[str]

# All venues at once
await client.live_symbols()                  # dict[exchange, list[str]]

# Symbols with persisted history
await client.historical_symbols("hl")        # list[str]

# Which venues quote ETH?
await client.instrument("ETH")               # InstrumentVenues

Historical data

History iterators paginate automatically — no cursor handling needed:

# Trades
async for trade in client.history_trades("hl", "ETH",
                                          start="2026-04-01",
                                          end="2026-04-08"):
    print(trade.time, trade.price, trade.size, trade.side)

# 5-second quote snapshots
async for snap in client.history_quotes("uni", "WETH/USDC",
                                         start="2026-04-01"):
    print(snap.time, snap.bids[0].price if snap.bids else None)

# Funding rates
async for fr in client.history_funding("hl", "ETH"):
    print(fr.ratePct, fr.intervalHrs)

# Yield-market snapshots
async for rate in client.history_rates("0xc62d...", start="2026-03-01"):
    print(rate.impliedApy, rate.tvl)

All methods accept start/end as ISO 8601 strings or epoch milliseconds. Default range: last 24 hours. Maximum lookback: 1 day (free tier) or 90 days (professional/api).

Account

await client.me()                  # JWT claims
await client.subscription_status() # current tier + expiry (re-reads on-chain)
await client.refresh_token()       # new 7-day JWT with updated tier claims

---

Message types

Type	Description	Venues
QuoteMessage	Top-of-book snapshot	All
PrintMessage	Trade execution	CLOB, AMM, Oracle*, Yield
FundingMessage	Perpetual funding rate	HL, dYdX, GMX, Vertex, Ostium
DepthMessage	Concentrated-liquidity tick snapshot + market impact	AMM
LiquidityMessage	LP Mint/Burn event	AMM
RateMarketMessage	Yield-market or lending-market snapshot	Pendle, Spectra, Aave, Compound, Morpho
RateDepthMessage	Depth-at-size APY quote	Pendle only
LendingActionMessage	On-chain lending event (Supply/Borrow/Liquidate/...)	Aave, Compound, Morpho
RateModelParamsMessage	IRM curve parameters	Aave, Compound, Morpho
AmmBookMessage	Consolidated AMM NBBO (professional+)	Virtual
AmmLiquiditySnapshotMessage	LP event backfill on subscribe (professional+)	Virtual
ArbFlagMessage	Cross-venue arb gap signal (super_admin)	Virtual
SpreadMessage	Same-venue cross-tier spread (professional+)	Uni, Univ4
FeedStaleMessage	Feed health: stale	Broadcast
FeedLiveMessage	Feed health: recovered	Broadcast
ServerClosingMessage	Graceful shutdown notice	Broadcast
ErrorMessage	Error frame	Server response

*GMX and Vertex do not emit PrintMessage (oracle settle on net positions).

Key field conventions

• time — always epoch milliseconds (int)
• side — 0 = aggressive sell (bid hit), 1 = aggressive buy (ask lift), 2 = unknown
• ratePct (FundingMessage) — annualized percentage, e.g. 10.95 = 10.95%/yr
• impliedApy (RateMarketMessage) — decimal fraction, e.g. 0.058 = 5.8%
• ptPrice — Pendle: USD price; Spectra: fraction of IBT
• tvl — Pendle: USD; Spectra v1: underlying-asset units (no USD oracle)
• amount0/amount1/amount (LiquidityMessage) — decimal strings in raw ERC-20 units

---

Tiers & limits

Tier	WS symbol cap	ammbook / spread	Historical data
Free (anonymous)	3	—	1 day
Professional	50	✓	90 days
API	100	✓	90 days

---

Error handling

from mackinac.exceptions import (
    AuthError,      # 401 / WS auth_error
    TierError,      # 403 / WS subscription_required
    RateLimitError, # 429 / WS rate_limited — has .retry_after
    SymbolLimitError,
    InvalidSymbolError,
    ServerError,    # 5xx / WS internal_error
)

try:
    async with client.subscribe("ammbook:WETH/USDC") as feed:
        ...
except TierError:
    print("Upgrade to professional tier for ammbook access")

• REST 4xx → typed exception immediately (no retry)
• REST 5xx → retried 3× with exponential backoff, then ServerError
• WS auth_error / subscription_required → raised, reconnect stops
• Other WS errors → delivered as ErrorMessage to the iterator

---

Installing extras

pip install 'mackinac-client[wallet]'   # EIP-191 wallet sign-in
pip install 'mackinac-client[tables]'   # pandas, for amm_fee_table + rate_table examples
pip install 'mackinac-client[dev]'      # testing + model codegen

---

Examples

Examples with both API flavours use a _sync / _async filename suffix. Older single-flavour examples keep their original names.

File	What it shows
examples/subscribe_quotes.py	HL ETH live top-of-book
examples/historical_trades.py	7-day Ostium XAU/USD → CSV
examples/yield_rates.py	Pendle + Spectra live APY table
examples/reconnect_pattern.py	Iterator survives forced disconnect
examples/multi_venue_basis.py	Cross-venue ETH basis across CLOB, AMM, and oracle venues
examples/amm_fee_table.py	AMM depth → pandas fee-tier impact table
examples/rate_table.py	rate_market stream → pandas DataFrame
examples/funding_carry.py	CME vs HL vs Ostium carry cost comparison
examples/lending_rates_sync.py	Aave + Compound + Morpho supply/borrow APY table (sync Mackinac)
examples/lending_rates_async.py	Same lending-rate table using AsyncClient
examples/liquidation_watch_sync.py	Live liquidation feed across all lending venues (sync)
examples/liquidation_watch_async.py	Same liquidation feed using AsyncClient

---

Versioning

mackinac.__version__ follows the library release. mackinac.__protocol_version__ matches the API schema version; a major bump means a breaking wire-format change.

---

License

MIT