Data Catalog

Browse our curated market data catalog across exchanges and asset types, and the raw trades, quotes and derivatives data behind it.

Preview data — no subscription. Try it free with DEMO-KEY

6 raw · 19 datasets

Raw data

The ticks the metrics are built from, as Parquet filesHow raw data works →

Raw L2 order books aren't offered yet. Ask us.

Trades

Derivatives

Market Data

L1 (Top of Book)

L2 (Order Book)

Preview Data

No account needed

A fixed slice of every dataset is free to download: the shared key DEMO-KEY over cURL, --preview in the CLI or preview=True in Python. Pick one for a command you can paste and run. A subscription unlocks full history and every exchange, symbol, interval and parameter; extended trials are available on request.

Raw data

The 2025-06 file of each venue's BTC perpetual, for every dataset the venue serves.

Dataset	Binanceperpetual-BTC-USDT:USDT	OKXperpetual-BTC-USDT:USDT	Hyperliquidperpetual-BTC-USDC:USDC
trades
quotes
mark_price
index_price
funding_rate
open_interest

Metrics

Every metric, over one slice: binance-futures · perpetual-BTC-USDT:USDT · 5m · exchange timestamps · 2025-05-01 → 2025-05-31

Trades

Derivatives

Market Data

L1 (Top of Book)

L2 (Order Book)

trades · binance-futures · perpetual-BTC-USDT:USDT · 2025-06

Open dataset

terminal

# Install the Aperiodic CLI into ~/.local/bin
mkdir -p "$HOME/.local/bin"
curl -fsSL https://raw.githubusercontent.com/aperiodic-io/cli/main/install.sh | INSTALL_DIR="$HOME/.local/bin" bash
export PATH="$HOME/.local/bin:$PATH"

# Download the 2025-06 trades file into ./raw — no API key needed
aperiodic raw trades --preview \
  --exchange binance-futures \
  --symbol perpetual-BTC-USDT:USDT \
  --output-dir ./raw && echo "Saved to $PWD/raw"

Mark price

Prime + Raw

The venue’s mark price, one row each time it changes.

/api/v1/data/raw/mark_price

Schema

exchange_timestamptimestampWhen the venue says the event happened (UTC, µs). Exact, except modelled on Hyperliquid, whose feed carries no exchange time: there the file has an exchange_timestamp_kind column ("modelled") right after it.

local_timestamptimestampWhen the event reached the capture machine (UTC, µs). Measured by our capture, or modelled as exchange_timestamp plus a latency drawn from our measured distribution; see local_timestamp_kind.

local_timestamp_kindstring"measured" or "modelled". Don't use modelled rows for latency research.

mark_pricefloat64Mark price

Coverage per venue

Live coverage is unavailable right now; every symbol's coverage is at /api/v1/metadata/raw.

File layout

One file per calendar month before 2026-08-01, one per day from then on, in the aperiodic-raw-derivatives bucket. Example key:

v1/mark_price/exchange=binance-futures/symbol=perpetual-BTC-USDT:USDT/year=2025/month=06/data.parquet

Try it

The shared DEMO-KEY gets the 2025-06 file of each venue's BTC perpetual. No account needed.

terminal

# Install the Aperiodic CLI into ~/.local/bin
mkdir -p "$HOME/.local/bin"
curl -fsSL https://raw.githubusercontent.com/aperiodic-io/cli/main/install.sh | INSTALL_DIR="$HOME/.local/bin" bash
export PATH="$HOME/.local/bin:$PATH"

# Download the 2025-06 mark_price file into ./raw — no API key needed
aperiodic raw mark_price --preview \
  --exchange binance-futures \
  --symbol perpetual-BTC-USDT:USDT \
  --output-dir ./raw && echo "Saved to $PWD/raw"

Code

pip install "aperiodic[polars]"

pypi ↗

example.py

from datetime import date
import aperiodic as ap

# Into one DataFrame (monthly files are trimmed to your range)
df = ap.get_raw(api_key="YOUR_KEY", dataset="mark_price",
                exchange="binance-futures", symbol="perpetual-BTC-USDT:USDT",
                start_date=date(2025, 6, 1), end_date=date(2025, 6, 30))
print(df.head())

# Or stream the files to disk, skipping ones you already have
ap.download_raw(api_key="YOUR_KEY", dataset="mark_price",
                exchange="binance-futures", symbol="perpetual-BTC-USDT:USDT",
                start_date=date(2025, 1, 1), end_date=date(2025, 12, 31),
                output_dir="raw")

Read the folder back one dataset at a time, with a glob:

terminal

# DuckDB, from the folder that holds ./raw
duckdb -c "SELECT * FROM read_parquet('raw/mark_price/**/data.parquet')
  WHERE exchange_timestamp >= '2025-06-01' LIMIT 10"

example.py

from datetime import datetime, timezone
import polars as pl

df = (
    pl.scan_parquet("raw/mark_price/**/data.parquet")
    .filter(pl.col("exchange_timestamp") >= datetime(2025, 6, 1, tzinfo=timezone.utc))
    .collect()
)
print(df.head())

Notes

exchange_timestamp is the venue's own time. local_timestamp is measured on days we captured ourselves and modelled from our measured latency before that; local_timestamp_kind says which. Don't use modelled days for latency research.
One row each time the value changes: the venue’s forward-filled ticker is undone, so rows are events, not samples.
On Hyperliquid, whose feed carries no exchange time, exchange_timestamp is modelled; those files say so in an exchange_timestamp_kind column.

Metrics built from this dataset

Data Catalog

Browse our curated market data catalog across exchanges and asset types, and the raw trades, quotes and derivatives data behind it.

Preview data — no subscription. Try it free with DEMO-KEY

6 raw · 19 datasets

Raw data

The ticks the metrics are built from, as Parquet filesHow raw data works →

Raw L2 order books aren't offered yet. Ask us.

Trades

Derivatives

Market Data

L1 (Top of Book)

L2 (Order Book)

Preview Data

No account needed

Raw data

The 2025-06 file of each venue's BTC perpetual, for every dataset the venue serves.

Dataset	Binanceperpetual-BTC-USDT:USDT	OKXperpetual-BTC-USDT:USDT	Hyperliquidperpetual-BTC-USDC:USDC
trades
quotes
mark_price
index_price
funding_rate
open_interest

Metrics

Every metric, over one slice: binance-futures · perpetual-BTC-USDT:USDT · 5m · exchange timestamps · 2025-05-01 → 2025-05-31

Trades

Derivatives

Market Data

L1 (Top of Book)

L2 (Order Book)

trades · binance-futures · perpetual-BTC-USDT:USDT · 2025-06

Open dataset

terminal

# Install the Aperiodic CLI into ~/.local/bin
mkdir -p "$HOME/.local/bin"
curl -fsSL https://raw.githubusercontent.com/aperiodic-io/cli/main/install.sh | INSTALL_DIR="$HOME/.local/bin" bash
export PATH="$HOME/.local/bin:$PATH"

# Download the 2025-06 trades file into ./raw — no API key needed
aperiodic raw trades --preview \
  --exchange binance-futures \
  --symbol perpetual-BTC-USDT:USDT \
  --output-dir ./raw && echo "Saved to $PWD/raw"

Mark price

Prime + Raw

The venue’s mark price, one row each time it changes.

/api/v1/data/raw/mark_price

Schema

local_timestamp_kindstring"measured" or "modelled". Don't use modelled rows for latency research.

mark_pricefloat64Mark price

Coverage per venue

Venue	Symbols	From
Binance	984	Mar 2020
OKX	700	Oct 2019
Hyperliquid	529	Mar 2025

File layout

One file per calendar month before 2026-08-01, one per day from then on, in the aperiodic-raw-derivatives bucket. Example key:

v1/mark_price/exchange=binance-futures/symbol=perpetual-BTC-USDT:USDT/year=2025/month=06/data.parquet

Try it

The shared DEMO-KEY gets the 2025-06 file of each venue's BTC perpetual. No account needed.

terminal

# Install the Aperiodic CLI into ~/.local/bin
mkdir -p "$HOME/.local/bin"
curl -fsSL https://raw.githubusercontent.com/aperiodic-io/cli/main/install.sh | INSTALL_DIR="$HOME/.local/bin" bash
export PATH="$HOME/.local/bin:$PATH"

# Download the 2025-06 mark_price file into ./raw — no API key needed
aperiodic raw mark_price --preview \
  --exchange binance-futures \
  --symbol perpetual-BTC-USDT:USDT \
  --output-dir ./raw && echo "Saved to $PWD/raw"

Code

pip install "aperiodic[polars]"

pypi ↗

example.py

from datetime import date
import aperiodic as ap

# Into one DataFrame (monthly files are trimmed to your range)
df = ap.get_raw(api_key="YOUR_KEY", dataset="mark_price",
                exchange="binance-futures", symbol="perpetual-BTC-USDT:USDT",
                start_date=date(2025, 6, 1), end_date=date(2025, 6, 30))
print(df.head())

# Or stream the files to disk, skipping ones you already have
ap.download_raw(api_key="YOUR_KEY", dataset="mark_price",
                exchange="binance-futures", symbol="perpetual-BTC-USDT:USDT",
                start_date=date(2025, 1, 1), end_date=date(2025, 12, 31),
                output_dir="raw")

Read the folder back one dataset at a time, with a glob:

terminal

# DuckDB, from the folder that holds ./raw
duckdb -c "SELECT * FROM read_parquet('raw/mark_price/**/data.parquet')
  WHERE exchange_timestamp >= '2025-06-01' LIMIT 10"

example.py

from datetime import datetime, timezone
import polars as pl

df = (
    pl.scan_parquet("raw/mark_price/**/data.parquet")
    .filter(pl.col("exchange_timestamp") >= datetime(2025, 6, 1, tzinfo=timezone.utc))
    .collect()
)
print(df.head())

Notes

exchange_timestamp is the venue's own time. local_timestamp is measured on days we captured ourselves and modelled from our measured latency before that; local_timestamp_kind says which. Don't use modelled days for latency research.
One row each time the value changes: the venue’s forward-filled ticker is undone, so rows are events, not samples.
On Hyperliquid, whose feed carries no exchange time, exchange_timestamp is modelled; those files say so in an exchange_timestamp_kind column.

Metrics built from this dataset

Mark price

Prime + Raw

The venue’s mark price, one row each time it changes.

/api/v1/data/raw/mark_price

Schema

local_timestamp_kindstring"measured" or "modelled". Don't use modelled rows for latency research.

mark_pricefloat64Mark price

Coverage per venue

Live coverage is unavailable right now; every symbol's coverage is at /api/v1/metadata/raw.

File layout

One file per calendar month before 2026-08-01, one per day from then on, in the aperiodic-raw-derivatives bucket. Example key:

v1/mark_price/exchange=binance-futures/symbol=perpetual-BTC-USDT:USDT/year=2025/month=06/data.parquet

Try it

The shared DEMO-KEY gets the 2025-06 file of each venue's BTC perpetual. No account needed.

terminal

# Install the Aperiodic CLI into ~/.local/bin
mkdir -p "$HOME/.local/bin"
curl -fsSL https://raw.githubusercontent.com/aperiodic-io/cli/main/install.sh | INSTALL_DIR="$HOME/.local/bin" bash
export PATH="$HOME/.local/bin:$PATH"

# Download the 2025-06 mark_price file into ./raw — no API key needed
aperiodic raw mark_price --preview \
  --exchange binance-futures \
  --symbol perpetual-BTC-USDT:USDT \
  --output-dir ./raw && echo "Saved to $PWD/raw"

Code

pip install "aperiodic[polars]"

pypi ↗

example.py

from datetime import date
import aperiodic as ap

# Into one DataFrame (monthly files are trimmed to your range)
df = ap.get_raw(api_key="YOUR_KEY", dataset="mark_price",
                exchange="binance-futures", symbol="perpetual-BTC-USDT:USDT",
                start_date=date(2025, 6, 1), end_date=date(2025, 6, 30))
print(df.head())

# Or stream the files to disk, skipping ones you already have
ap.download_raw(api_key="YOUR_KEY", dataset="mark_price",
                exchange="binance-futures", symbol="perpetual-BTC-USDT:USDT",
                start_date=date(2025, 1, 1), end_date=date(2025, 12, 31),
                output_dir="raw")

Read the folder back one dataset at a time, with a glob:

terminal

# DuckDB, from the folder that holds ./raw
duckdb -c "SELECT * FROM read_parquet('raw/mark_price/**/data.parquet')
  WHERE exchange_timestamp >= '2025-06-01' LIMIT 10"

example.py

from datetime import datetime, timezone
import polars as pl

df = (
    pl.scan_parquet("raw/mark_price/**/data.parquet")
    .filter(pl.col("exchange_timestamp") >= datetime(2025, 6, 1, tzinfo=timezone.utc))
    .collect()
)
print(df.head())

Notes

exchange_timestamp is the venue's own time. local_timestamp is measured on days we captured ourselves and modelled from our measured latency before that; local_timestamp_kind says which. Don't use modelled days for latency research.
One row each time the value changes: the venue’s forward-filled ticker is undone, so rows are events, not samples.
On Hyperliquid, whose feed carries no exchange time, exchange_timestamp is modelled; those files say so in an exchange_timestamp_kind column.

Metrics built from this dataset

Mark price

Prime + Raw

The venue’s mark price, one row each time it changes.

/api/v1/data/raw/mark_price

Schema

local_timestamp_kindstring"measured" or "modelled". Don't use modelled rows for latency research.

mark_pricefloat64Mark price

Coverage per venue

Venue	Symbols	From
Binance	984	Mar 2020
OKX	700	Oct 2019
Hyperliquid	529	Mar 2025

File layout

One file per calendar month before 2026-08-01, one per day from then on, in the aperiodic-raw-derivatives bucket. Example key:

v1/mark_price/exchange=binance-futures/symbol=perpetual-BTC-USDT:USDT/year=2025/month=06/data.parquet

Try it

The shared DEMO-KEY gets the 2025-06 file of each venue's BTC perpetual. No account needed.

terminal

# Install the Aperiodic CLI into ~/.local/bin
mkdir -p "$HOME/.local/bin"
curl -fsSL https://raw.githubusercontent.com/aperiodic-io/cli/main/install.sh | INSTALL_DIR="$HOME/.local/bin" bash
export PATH="$HOME/.local/bin:$PATH"

# Download the 2025-06 mark_price file into ./raw — no API key needed
aperiodic raw mark_price --preview \
  --exchange binance-futures \
  --symbol perpetual-BTC-USDT:USDT \
  --output-dir ./raw && echo "Saved to $PWD/raw"

Code

pip install "aperiodic[polars]"

pypi ↗

example.py

from datetime import date
import aperiodic as ap

# Into one DataFrame (monthly files are trimmed to your range)
df = ap.get_raw(api_key="YOUR_KEY", dataset="mark_price",
                exchange="binance-futures", symbol="perpetual-BTC-USDT:USDT",
                start_date=date(2025, 6, 1), end_date=date(2025, 6, 30))
print(df.head())

# Or stream the files to disk, skipping ones you already have
ap.download_raw(api_key="YOUR_KEY", dataset="mark_price",
                exchange="binance-futures", symbol="perpetual-BTC-USDT:USDT",
                start_date=date(2025, 1, 1), end_date=date(2025, 12, 31),
                output_dir="raw")

Read the folder back one dataset at a time, with a glob:

terminal

# DuckDB, from the folder that holds ./raw
duckdb -c "SELECT * FROM read_parquet('raw/mark_price/**/data.parquet')
  WHERE exchange_timestamp >= '2025-06-01' LIMIT 10"

example.py

from datetime import datetime, timezone
import polars as pl

df = (
    pl.scan_parquet("raw/mark_price/**/data.parquet")
    .filter(pl.col("exchange_timestamp") >= datetime(2025, 6, 1, tzinfo=timezone.utc))
    .collect()
)
print(df.head())

Notes

exchange_timestamp is the venue's own time. local_timestamp is measured on days we captured ourselves and modelled from our measured latency before that; local_timestamp_kind says which. Don't use modelled days for latency research.
One row each time the value changes: the venue’s forward-filled ticker is undone, so rows are events, not samples.
On Hyperliquid, whose feed carries no exchange time, exchange_timestamp is modelled; those files say so in an exchange_timestamp_kind column.