Arithmax Chest

Normalized market data. One API. Every provider.
Equities · Crypto · Futures · Macro

---

Arithmax Chest gives you a single, normalized interface to fetch OHLCV market data across multiple asset classes — without worrying about which provider powers each symbol. Just ask for the data and the backend handles the routing.

Quick Install

pip install arithmaxchest

For optional providers:

pip install arithmaxchest[all]        # everything (yfinance + databento + uvicorn)
pip install arithmaxchest[yahoo]      # equities via yfinance
pip install arithmaxchest[futures]    # futures via databento
pip install arithmaxchest[server]     # self-host the API

Complete User Guide

1. Create a client

from achest import MarketDataClient

# Hosted API (defaults to https://achest.misango.me)
client = MarketDataClient()

# Or with your own server + token
client = MarketDataClient(host="https://your-server.com", token="your-api-token")

Always use as a context manager:

with MarketDataClient() as client:
    data = client.get(...)

2. The get() method — every parameter explained

client.get(
    symbols,          # list of ticker strings
    start,            # start date — see date formats below
    end,              # end date — see date formats below
    resolution,       # bar size — see resolutions below
    provider          # force a specific provider or "auto"
)

symbols — what tickers look like

Asset Class	Example Symbols	Notes
Equities	"AAPL", "MSFT", "TSLA", "SPY"	Standard tickers, 1–5 letters
Crypto	"BTCUSDT", "ETHUSDT", "SOLUSDT", "BTC-USD"	Ends in USDT, USDC, or -USD
Futures	"ES.FUT", "NQ.FUT", "CL.FUT"	Ends in .FUT (continuous contracts)
Indices	"^GSPC", "^VIX", "^DJI"	Starts with ^
Economic	"GDP", "UNRATE", "CPIAUCSL", "FEDFUNDS"	Predefined macro symbols

You can mix asset classes in one call:

client.get(["AAPL", "BTCUSDT", "ES.FUT", "^GSPC"], "2025-01-01", "2025-01-31")

start / end — date formats accepted

All of these work:

# ISO strings
client.get(["AAPL"], "2025-01-01", "2025-01-31")

# datetime.date objects
from datetime import date
client.get(["AAPL"], date(2025, 1, 1), date(2025, 1, 31))

# datetime.datetime objects
from datetime import datetime
client.get(["AAPL"], datetime(2025, 1, 1), datetime(2025, 1, 31))

resolution — bar sizes

Resolution	Description	Example
"tick"	Trade-by-trade (raw ticks)	client.get(["BTCUSDT"], ..., resolution="tick")
"second"	1-second bars	client.get(["AAPL"], ..., resolution="second")
"minute"	1-minute bars	client.get(["AAPL"], ..., resolution="minute")
"hour"	1-hour bars	client.get(["AAPL"], ..., resolution="hour")
"daily"	1-day bars (default)	client.get(["AAPL"], ..., resolution="daily")

# Minute-level data for day trading
client.get(["AAPL"], "2025-01-02", "2025-01-03", resolution="minute")

# Hourly data for swing trading
client.get(["BTCUSDT"], "2025-01-01", "2025-01-07", resolution="hour")

provider — auto vs. manual override

Value	Behavior
"auto" (default)	Backend selects the best available provider based on symbol + resolution + configured API keys
"yahoo"	Force Yahoo Finance
"binance"	Force Binance (crypto only)
"databento"	Force Databento (futures, equities)
"alpaca"	Force Alpaca
"tiingo"	Force Tiingo
"alpha_vantage"	Force Alpha Vantage
"fred"	Force FRED (economic data only)
"quandl"	Force Quandl
"massive"	Force Massive

# Auto-select (recommended)
client.get(["ES.FUT"], "2025-01-01", "2025-01-31", provider="auto")

# Force a specific provider
client.get(["AAPL"], "2025-01-01", "2025-01-31", provider="yahoo")

3. Full examples

Equities — daily data

from achest import MarketDataClient

with MarketDataClient() as client:
    aapl = client.get(["AAPL", "MSFT", "GOOGL"], "2025-01-01", "2025-01-31")
    print(aapl.head())

Crypto — minute data

with MarketDataClient() as client:
    btc = client.get(["BTCUSDT"], "2025-01-15", "2025-01-16", resolution="minute")
    print(btc.head())

Futures — daily via Databento

with MarketDataClient() as client:
    es = client.get(["ES.FUT", "NQ.FUT"], "2025-01-01", "2025-01-31")
    print(es.head())

Mixed assets in one call

with MarketDataClient() as client:
    data = client.get(["AAPL", "BTCUSDT", "ES.FUT"], "2025-01-02", "2025-01-10")
    print(data.groupby("symbol").tail(3))

4. Download to file

# CSV
client.get(["AAPL"], "2025-01-01", "2025-01-31", format="csv", output="aapl.csv")

# Parquet (fast, compressed)
client.get(["AAPL"], "2025-01-01", "2025-01-31", format="parquet", output="aapl.parquet")

# JSON
client.get(["AAPL"], "2025-01-01", "2025-01-31", format="json", output="aapl.json")

5. The download() method

client.download(["SPY", "QQQ"], "2025-01-01", "2025-01-31",
                output="market-data.parquet")

Data Shape

All responses return normalized OHLCV data:

timestamp	symbol	provider	open	high	low	close	volume
2025-01-02	AAPL	yahoo	243.5	245.8	242.3	244.9	48234500
2025-01-02	BTCUSDT	binance	94250	95800	93800	95430	12500
2025-01-02	ES.FUT	databento	5939.25	5995.25	5874.75	5914.75	1714061

Every row has:

• timestamp — UTC-normalized datetime
• symbol — the original ticker you requested
• provider — which provider served the data
• open, high, low, close, volume — OHLCV values

Symbol Reference

Input	Classified As	Routed To
AAPL	equity	yahoo (default)
MSFT	equity	yahoo (default)
BTCUSDT	crypto	binance (if key exists) or yahoo
ETHUSDT	crypto	binance (if key exists) or yahoo
ES.FUT	futures	databento (if key exists) or massive
NQ.FUT	futures	databento (if key exists) or massive
CL.FUT	futures	databento (if key exists) or massive
^GSPC	index	yahoo
^VIX	index	yahoo
GDP	economic	fred (if key exists) or quandl
UNRATE	economic	fred (if key exists) or quandl

Self-Hosting

Arithmax Chest is also a FastAPI server you can deploy yourself:

pip install arithmaxchest[all]
uvicorn achest.server:app --host 0.0.0.0 --port 8000

With Docker and EC2 support — see the documentation for deployment details.

Why Arithmax Chest?

• Normalized schema — Every symbol returns the same columns regardless of provider
• Provider abstraction — The backend picks the best source; your code never changes
• Multi-asset — Equities, crypto, futures, indices, economic indicators — one API
• Research-ready — Clean DataFrames with no provider-specific wrangling
• Portable — Works in notebooks, scripts, pipelines, and production systems

License

MIT

---

Built by Arithmax Research. Market data for the next generation of analysts and algorithms.