Pre-registration open

The hub for crypto market data.

Spot, perpetuals, funding and full-depth order books from every major exchange. L2, L3 and L4, normalized into one schema. Historical and live. Pre-register now for early access and pre-sale pricing.

        EXCHANGES ↓ ONE SCHEMA
        SYNTHETIC PREVIEW
      

StatusPre-registration
First venueHyperliquid
Schema1
DepthL2–L4

Every dataset a trading desk needs, in one place.

Buy what you need, when you need it. Every dataset will be available as a historical archive and as a live stream.

Browse the dataset catalog

SPOTall major venues

Spot markets

Trades, quotes and candles for every spot pair, with exchange and capture timestamps on every message.

HistoricalLive

PERPperps · futures · options

Derivatives

Perpetuals, dated futures and options, including liquidations, mark and index prices.

HistoricalLive

FUNDevery funding interval

Funding & open interest

Funding rates and open interest across venues, aligned on one clock so they compare directly.

Cross-venueLive

L2/L3/L4full depth

Order book depth

Aggregated levels, order-by-order books and attributed order flow. Rebuild any book at any moment.

L2L3L4

The fields with real statistical value that nobody calculates.

Beyond raw trades and books, our data is extended with derived fields computed from full-depth data on a single clock. Ready to use, so you don't have to rebuild them.

See every field and how it's computed

          OFI-ML
          from L2
        

Multi-level order flow imbalance Net pressure across the top levels of the book, not just the best bid and ask.

          VPIN
          from trades
        

Flow toxicity Volume-synchronized probability of informed trading, per instrument.

          QUEUE
          from L3
        

Queue dynamics Order lifetime, queue position and fill probability, reconstructed from order-level data.

          CXL/FILL
          from L3
        

Cancel-to-fill ratio How much displayed liquidity is real, by price level and over time.

          RESIL
          from L2
        

Book resiliency How fast depth refills after a sweep, measured in time and size.

          LAMBDA
          from trades
        

Price impact Rolling estimate of price move per unit of signed volume.

          MARKOUT
          from trades
        

Markouts Post-trade price drift at fixed horizons, split by aggressor side.

          PART
          from L4
        

Participant flow Aggregated behaviour by attributed participant, where the venue publishes it.

Backed by empirical research. Proven to have better predictive power.*

Every extended field is tested before it ships. In our studies, models given extended fields consistently outperformed the same models trained on raw trades and top-of-book data alone.

Read the study

Baseline · raw trades + L151.6%
+ FDC extended fields54.9%
Directional hit rate, out of sample+3.3 pts

* Internal research. Out of sample on BTC, ETH and SOL perpetuals (Hyperliquid), predicting the sign of the mid-price return over the next 500 ms, sampled every 100 ms. Trained Jan–Jun 2026, tested Jul–Sep 2026, walk-forward with no overlap. Baseline and test use the same gradient-boosted model and hyperparameters; only the feature set differs. Results exclude fees and latency, may not generalize to other assets, venues or horizons, and do not constitute investment advice.

From aggregated levels to individual orders.

How L2, L3 and L4 differ

FUNDING · PERPETUALSCROSS-VENUE

negative funding · OI · mark · index positive

          PYTHON
          REST
          WEBSOCKET
          S3
        

CSV · PARQUET · JSON○ backfilling…

Follow the same trader across every venue.

Our blockchain intelligence links on-chain activity to exchange order flow. We cluster wallets into entities, match deposits and withdrawals to venues, and track trades as they move across exchanges. The result is L4+ data: order flow with cross-venue attribution.

01Entity clusteringAddresses grouped into entities using on-chain behaviour and transaction patterns.

02Venue attributionDeposits and withdrawals matched to exchanges and to the order flow that follows.

03Cross-venue trade trackingLinked activity across exchanges, with a confidence score on every match.

How attribution works

          ENTITY TRACE · E-4417SYNTHETIC
        
          match confidencetracing hop 1 / 5

All major exchanges, added in phases.

We start with Hyperliquid and add exchanges in phases as each one is captured and normalized to our standard. L3 and L4 are offered wherever a venue publishes order-level data. Need a venue that isn't listed? Tell us and we'll prioritize it.

Full coverage by venue

        ROADMAP · EXCHANGE AVAILABILITY →
        TARGETS · SUBJECT TO CHANGE
      

Built by people who were tired of cleaning data.

Finance Data Cloud is a five-person startup building a single source for crypto market data. Every exchange publishes data differently: different symbols, timestamps, book formats and gaps. We capture it at the source, normalize it into one schema, and make it available as both a historical archive and a live stream. Today we're building the capture and normalization pipeline and opening pre-registration ahead of our first release.

We're building for quant funds, market makers, researchers and builders who would rather spend their time on models than on parsers.

More about the company

01 Captured at the source Raw messages are recorded directly from each exchange, with capture timestamps, before anything is transformed.

02 One schema Symbols, sides, timestamps and book updates follow the same format on every venue.

03 Historical and live The archive and the live stream share the same fields, so research code runs in production unchanged.

Five people. One pipeline.

Meet the team We're hiring

portraitDG

Daniel Garcia CEO & Founder

Spent years working with trading desks and market-structure research. Founded Finance Data Cloud to fix the data problems he kept running into.

portraitCS

Chris Smith Co-founder

Has built data platforms for fintech and exchange infrastructure teams. Leads product, partnerships and the commercial side.

portraitMO

Maya Okafor Data Engineer

Owns the normalization layer: the schema that makes a Binance trade and a Deribit trade look the same.

portraitLB

Lukas Brandt Backend Engineer

Builds the APIs and streaming services, with a focus on low-latency delivery and gap-free replay.

portraitPR

Priya Raman Infrastructure Engineer

Runs capture infrastructure across regions so no message is missed, from collectors to long-term storage.

Pre-register. Talk to us first.

Pre-registered partners get pre-sale pricing and a say in which exchanges and datasets ship first. Reach our founder directly.

DANIEL GARCIA · CEO & FOUNDERdaniel@financedatacloud.com → EARLY ACCESS · PRE-SALE PRICINGPre-register with your requirements →