Skip to content

Market Data Provider Research: US & Hong Kong Stocks / Options Data Comparison

Version: 2026-10-14
Language: English

Prepared for: Options / Market Data Platform Research
Date: 14 October 2026
Status: Research / Architecture Planning
Primary Goal: Compare market-data providers for US and Hong Kong equities and options, including pricing, historical coverage, real-time data, corporate actions, licensing, and suitability for backtesting or commercial applications.

Important: Pricing, exchange entitlements, and licensing terms can change. Always verify the latest provider and exchange terms before subscribing or using data in a commercial/customer-facing system.


For a part-time development project focused on options research, strategy calculation, screening, and backtesting, the recommended approach is:

Use free plans first:

  • Massive (formerly Polygon.io) — US stocks / US options
  • EODHD — global stocks / Hong Kong stocks / dividends / splits / fundamentals

Approximate cost:

US$0/month

The goal is to build the software architecture before paying for large amounts of market data.


Suggested combination:

  • Massive Options Starter
  • EODHD paid plan if required

Approximate cost:

US$50–60/month

Use this stage for:

  • option chain testing
  • Greeks
  • implied volatility
  • open interest
  • historical candles
  • corporate actions
  • strategy calculation
  • first version of the backtest engine

Upgrade to:

  • Massive Options Developer
  • EODHD for stock / global / HK data

Approximate cost:

~US$100/month

The purpose is to obtain deeper historical options data and perform realistic strategy backtests.


Possible options:

  • Massive Advanced
  • Intrinio
  • Databento
  • FirstRate Data historical datasets

Approximate cost:

US$200–350+/month, depending on data requirements.


If the application becomes:

  • multi-user
  • customer-facing
  • subscription-based
  • used by students/clients
  • redistributing exchange data

then personal plans are generally not sufficient.

Commercial/business licensing must be investigated.

For Hong Kong, exchange licensing can become significantly more expensive than US market data.


The most practical approach is:

US market first
↓
Build the platform
↓
Validate calculations
↓
Build backtesting
↓
Stabilise architecture
↓
Add Hong Kong stocks
↓
Investigate Hong Kong options licensing

The main reason is not programming complexity.

The main constraints are:

  1. market-data licensing
  2. exchange fees
  3. historical options availability
  4. redistribution rights
  5. commercial-use restrictions

US market data is generally easier and cheaper to obtain.

Hong Kong options data is substantially harder and more expensive.


The system should treat market data as several separate categories.

Required fields may include:

  • symbol
  • exchange
  • trading date
  • timestamp
  • open
  • high
  • low
  • close
  • adjusted close
  • volume
  • VWAP
  • number of trades

Timeframes may include:

  • daily
  • hourly
  • 30-minute
  • 15-minute
  • 5-minute
  • 1-minute
  • tick-level

Required corporate-action data includes:

  • cash dividends
  • special dividends
  • stock splits
  • reverse splits
  • spin-offs
  • symbol changes
  • mergers
  • delistings

For historical backtesting, corporate actions are essential because raw historical prices can otherwise produce misleading results.


Possible fields include:

  • market capitalisation
  • revenue
  • EPS
  • P/E
  • debt
  • cash flow
  • ROE
  • profit margin
  • earnings dates
  • analyst estimates

This is useful for scanners but is not essential for the first version of an options strategy engine.


For each option contract:

Underlying
Expiry
Strike
Call / Put
Bid
Ask
Last
Volume
Open Interest
Implied Volatility
Delta
Gamma
Theta
Vega
Rho
Contract Multiplier
Exercise Style
Settlement Type

For more advanced systems:

Bid Size
Ask Size
Trade Timestamp
Quote Timestamp
NBBO
IV Surface
Historical Greeks
Historical Bid / Ask
Historical Open Interest
Historical Option Chain

Data TypeScannerBacktestOptions EnginePriority
Daily OHLCVYesYesYes★★★★★
Intraday OHLCVYesYesYes★★★★★
Adjusted PricesYesYesNo★★★★★
DividendsNoYesYes★★★★★
Stock SplitsNoYesYes★★★★★
EarningsYesYesYes★★★★
Delisted StocksYesYesNo★★★★
Option ChainNoYesYes★★★★★
Bid / AskNoYesYes★★★★★
VolumeNoYesYes★★★★★
Open InterestNoYesYes★★★★★
Implied VolatilityNoYesYes★★★★★
GreeksNoYesYes★★★★★
Historical Option ChainNoYesYes★★★★★
Historical Bid / AskNoYesYes★★★★★
Tick DataAdvancedAdvancedAdvanced★★★
Level 2 / Order BookAdvancedRareRare★★

ProviderUS StocksHK StocksUS OptionsHK OptionsDividend / SplitHistorical OptionsTypical Entry PriceBest Use
Massive★★★★★No★★★★★NoYes★★★★★Free / 29/29 / 79 / $199+US options
EODHD★★★★Yes★★★★NoYes★★★★Free / ~$20–100Global stocks
Twelve Data★★★★YesLimitedNoYesLimitedFree / 79/79 / 229+Global market API
Alpha Vantage★★★★Global★★★★NoYes★★★★Free / PremiumResearch / prototype
Intrinio★★★★★Mainly US★★★★★NoYes★★★★★~$150+Professional options research
FMP★★★★Global on higher plansLimitedNoYesLimitedFree / 19/19 / 49 / $99+Fundamentals
Tiingo★★★★LimitedLimitedNoYesLimitedFree / ~$30+Stocks / internal research
AllTickYes★★★★★LimitedUnclearLimitedLimited~$99–399 USDTHK real-time
Databento★★★★★No★★★★★NoLimited★★★★★Usage basedInstitutional / quant
FirstRate Data★★★★No★★★★★NoLimited★★★★★Dataset / ~$79 updateHistorical backtest
IBKR API★★★★★★★★★★★★Depends on entitlementLimitedLimitedExchange subscriptionsExecution / live
Webull API★★★★★★★★US OptionsLimitedLimitedLimitedSubscription / entitlementBroker-linked data
HKEXHK Official★★★★★N/A★★★★★Official★★★★★HighOfficial HK data

Formerly known as Polygon.io.

Website:

https://massive.com/

Massive is one of the strongest choices for US options development.

Coverage includes:

  • US equities
  • US options
  • historical aggregates
  • trades
  • quotes
  • snapshots
  • option chains
  • Greeks
  • implied volatility
  • open interest
  • WebSocket feeds

Approximate price:

US$0/month

Typical features:

  • US option contracts
  • reference data
  • limited historical data
  • end-of-day data
  • minute aggregates
  • corporate actions
  • API rate limits

Best use:

  • prototype
  • schema development
  • early strategy calculation

Approximate price:

US$29/month

Typical additions:

  • unlimited API calls
  • delayed market data
  • Greeks
  • IV
  • open interest
  • snapshots
  • WebSocket
  • second aggregates
  • flat files

Best use:

  • development
  • live testing
  • option scanner

Approximate price:

US$79/month

Main advantage:

  • deeper historical options coverage
  • trades
  • more suitable for backtesting

This is likely the first paid plan worth using when the backtest engine becomes operational.


Approximate price:

US$199/month

Typical additions:

  • real-time
  • deeper historical access
  • quotes
  • trades

Best use:

  • serious research
  • live analytics

This is important.

Low-cost plans are generally intended for individual/non-professional use.

A customer-facing application may require:

  • business licensing
  • exchange licensing
  • display rights
  • redistribution rights

Do not assume that an individual US29orUS29 or US79 subscription allows raw market data to be displayed to clients.


Website:

https://eodhd.com/

EODHD is particularly useful for global stock coverage.

Coverage includes:

  • global equities
  • Hong Kong stocks
  • EOD prices
  • intraday prices
  • dividends
  • splits
  • fundamentals
  • earnings
  • news
  • technical indicators
  • US options

Typical features:

  • limited API calls
  • recent historical data
  • EOD
  • dividends
  • splits

Best use:

  • prototype
  • proof of concept

Higher plans may include:

  • 30+ years of historical data
  • intraday data
  • fundamentals
  • delayed live data
  • stock screening
  • larger API allowance

All-in-One pricing has historically been around:

US$99.99/month

Pricing should be checked before subscription.


EODHD options data may include:

  • OHLC
  • bid
  • ask
  • bid size
  • ask size
  • volume
  • open interest
  • open-interest changes
  • implied volatility
  • delta
  • gamma
  • theta
  • vega
  • rho
  • theoretical price
  • moneyness
  • DTE

EODHD can therefore potentially reduce the number of providers needed in a smaller system.


Website:

https://intrinio.com/

Intrinio is more expensive but more suitable for serious options analytics.

Approximate starting price:

US$150/month

Potential coverage:

  • US fundamentals
  • historical stocks
  • options history
  • tick-level data
  • Greeks
  • IV
  • OptionsEdge products

Commercial plans can start substantially higher.

Commercial plans may include:

  • commercial use
  • display rights
  • larger API limits
  • multiple users

This becomes attractive only after the software has proven commercial value.


Possible fields include:

  • closing option price
  • closing bid
  • closing ask
  • volume
  • open interest
  • IV
  • Greeks
  • full chains
  • volatility surfaces

Historical bulk datasets may extend much further back than standard API products.


Website:

https://www.alphavantage.co/

Stocks:

  • daily
  • weekly
  • monthly
  • intraday
  • adjusted / unadjusted
  • dividends
  • splits

Options:

  • current options chain
  • historical options
  • Greeks
  • implied volatility

Alpha Vantage is useful for:

  • research
  • prototype
  • long-term historical experiments

For commercial use, separate licensing should be confirmed.


Website:

https://twelvedata.com/

Twelve Data is strong for global market coverage.

Free

Good for:

  • development
  • small API usage
  • US equities
  • forex
  • crypto

Approximate price:

US$79/month

Possible features:

  • global markets
  • US equities
  • EOD international equities
  • fundamentals
  • dividends
  • splits
  • earnings

Approximate price:

US$229/month

Additional:

  • more global exchanges
  • higher limits
  • more real-time/delayed markets

Approximate price:

US$999/month

Suitable for higher-volume usage.


Individual plans are generally intended for personal/internal non-commercial usage.

External distribution normally requires business licensing.


Website:

https://financialmodelingprep.com/

FMP is especially strong for fundamental and company financial data.

PlanApproximate Monthly Cost
BasicFree
Starter~$19
Premium~$49
Ultimate~$99

Possible data:

  • financial statements
  • ratios
  • estimates
  • earnings
  • historical prices
  • ETF data
  • 13F data
  • transcripts
  • global equities

FMP is better suited to:

  • stock screening
  • fundamentals
  • company research

It is not the first choice for a dedicated options engine.


Website:

https://www.tiingo.com/

Typical strengths:

  • US equities
  • long historical coverage
  • adjusted prices
  • research-friendly API

Approximate paid price:

~US$30/month

Important:

Some Tiingo plans are intended for internal use only.

External sharing or redistribution may require separate licensing.


Website:

https://alltick.co/

AllTick is interesting for Hong Kong real-time data.

Approximate price:

99 USDT/month

May include:

  • selected products
  • real-time ticks
  • order book
  • K-lines

Approximate price:

199 USDT/month

Higher symbol limits.

Approximate price:

399 USDT/month

Larger market coverage.

Approximate price:

399 USDT/month

Potentially includes thousands of HK symbols with:

  • real-time quotes
  • order book
  • ticks
  • K-lines

The technical availability of data does not automatically mean commercial redistribution is permitted.

Before using AllTick in a commercial system, confirm:

  • commercial use
  • display rights
  • redistribution rights
  • exchange-authorised status
  • derived-data rules

Website:

https://databento.com/

Databento is designed for quantitative and institutional workloads.

  • tick data
  • trades
  • quotes
  • NBBO
  • OPRA options
  • high-quality historical data
  • exchange-native data schemas

Databento pricing often consists of:

Data Usage
+
Exchange Licence
+
Professional / Non-professional Entitlements

This differs from the simple fixed-price SaaS model.


Databento is suitable for:

  • quantitative research
  • professional backtesting
  • microstructure research
  • tick-level simulations
  • institutional-grade datasets

It may be unnecessary for the first version of this project.


Website:

https://firstratedata.com/

FirstRate Data is particularly useful when the main goal is:

historical backtesting

instead of real-time feeds.

  • historical option chains
  • bid
  • ask
  • open interest
  • implied volatility
  • Greeks
  • volume
  • historical underlyings

Coverage can include:

  • active symbols
  • delisted symbols
  • older historical periods

Instead of paying for every API request:

Buy / Subscribe to Dataset
↓
Download
↓
Store Locally
↓
Convert to Parquet
↓
Backtest Locally

This approach is excellent for large-scale backtesting.


Website:

https://www.interactivebrokers.com/

IBKR is excellent for:

  • brokerage
  • execution
  • live validation
  • real-time account positions
  • trading automation

However, it is not ideal as the primary long-term historical database.


Massive / EODHD / Dataset Provider
↓
Research
↓
Backtesting
↓
Strategy Signal
↓
IBKR
↓
Execution

Use IBKR as the execution layer rather than the core historical-data warehouse.


Broker APIs may expose:

  • US stocks
  • HK stocks
  • US options
  • real-time quotes
  • delayed quotes
  • account information
  • orders

However:

  • historical depth may be limited
  • rate limits may exist
  • redistribution rights may be restrictive
  • market data may require account entitlements
  • the API may be designed for trading rather than bulk research

These APIs are better suited to:

Live Trading
Execution
Account State
Current Quotes

rather than building a large historical research database.


Hong Kong market data should be treated separately from US data.

The main challenge is:

licensing

rather than software development.


Website:

https://www.hkex.com.hk/

HKEX is the official source for Hong Kong market data.

Possible data categories include:

  • equities
  • derivatives
  • stock options
  • futures
  • trades
  • quotes
  • order books
  • historical datasets

Commercial redistribution may involve large exchange licence fees.

Example licence structures can include quarterly fees for:

  • Securities Standard
  • Securities Premium
  • Securities FullTick
  • delayed data

These can run into tens of thousands of HKD per quarter.

Additional costs may include:

  • connection fees
  • subscriber fees
  • terminal fees
  • vendor fees
  • redistribution fees

HKEX provides official historical datasets.

Potential products include:

  • EOD data
  • transaction data
  • full-book data
  • derivatives data
  • options data

Historical datasets may be sold separately from current subscriptions.

This distinction is important:

Current Market Subscription
≠
Historical Archive Purchase

Historical full-market datasets can become expensive.


Many inexpensive data APIs support:

US Stocks ✓
US Options ✓
HK Stocks ✓
HK Options ✗

The reason is often exchange licensing.

Hong Kong options can involve:

  • HKEX licensing
  • specialised vendors
  • commercial display rights
  • per-user fees
  • redistribution restrictions

Therefore, Hong Kong options should be introduced only after the platform is stable.


The system should distinguish four levels of use.


Example:

Harry runs Python locally
↓
Downloads data
↓
Performs research
↓
No one else sees raw market data

Many individual plans are suitable.


Example:

Harry
+
Project Partner
+
Small Internal Team

This may already exceed a strict “individual use” licence.

Provider permission should be confirmed.


Example:

User logs in
↓
Selects stock
↓
System displays option chain
↓
Shows bid / ask / Greeks / IV

This is commercial display.

Business or exchange licensing may be required.


Example API output:

{
"symbol": "AAPL",
"expiry": "2026-11-20",
"strike": 250,
"type": "CALL",
"bid": 5.20,
"ask": 5.35,
"iv": 0.312,
"delta": 0.42
}

If this raw or near-raw market data is delivered to external users, it may qualify as redistribution.

This is the most restrictive licensing category.


A very important design strategy is to separate raw market data from derived analytics.

Example:

Raw Market Data
↓
Calculation Engine
↓
Strategy Analytics
↓
Derived Result

Instead of displaying:

Bid = $5.20
Ask = $5.35
IV = 31.2%
Delta = 0.42

the application might display:

Strategy: Bull Put Spread
Short Put: 200
Long Put: 190
Max Profit: $430
Max Loss: $570
Breakeven: $195.70
Probability Estimate: 68%

Derived-data licensing can sometimes differ from raw market-data redistribution.

However:

Never assume derived data is automatically unrestricted.

The provider and exchange terms must still be checked.


flowchart TD
A[Market Data Providers]
A --> B1[Massive]
A --> B2[EODHD]
A --> B3[HK Provider]
A --> B4[Historical Dataset]
B1 --> C[Provider Adapter Layer]
B2 --> C
B3 --> C
B4 --> C
C --> D[Normalized Market Data Schema]
D --> E1[Stock Market Database]
D --> E2[Options Database]
D --> E3[Corporate Actions Database]
D --> E4[Fundamental Database]
E1 --> F[Calculation Engine]
E2 --> F
E3 --> F
E4 --> F
F --> G1[Strategy Engine]
F --> G2[Backtest Engine]
F --> G3[Scanner]
F --> G4[Risk Engine]
G1 --> H[Result API]
G2 --> H
G3 --> H
G4 --> H
H --> I1[Web UI]
H --> I2[Reports]
H --> I3[Trading Signal]
H --> I4[Broker Integration]

A simple conceptual model:

flowchart LR
A[Data Layer] --> B[Calculation Layer]
B --> C[Output Layer]
A --> A1[Stocks]
A --> A2[Options]
A --> A3[Corporate Actions]
B --> B1[Greeks]
B --> B2[Strategies]
B --> B3[Risk]
B --> B4[Backtest]
C --> C1[UI]
C --> C2[Reports]
C --> C3[Alerts]
C --> C4[Broker Orders]

28. Backtesting Should Be a Separate Module

Section titled “28. Backtesting Should Be a Separate Module”

The platform should have four major independent systems:

flowchart TD
A[Data Collection] --> B[Data Storage]
B --> C[Calculation Engine]
C --> D[Strategy Engine]
C --> E[Backtest Engine]
C --> F[Scanner Engine]
D --> G[Output]
E --> G
F --> G

Backtesting should not be mixed directly with live strategy calculation.


Do not scatter vendor-specific code across the system.

Bad design:

polygon.get_options()
polygon.get_candles()
eodhd.get_dividend()

across many parts of the application.

Instead define a common interface.

Example:

class MarketDataProvider:
def get_candles(self, symbol, start, end, interval):
pass
def get_dividends(self, symbol, start, end):
pass
def get_splits(self, symbol, start, end):
pass
def get_option_chain(self, symbol, expiry=None):
pass
def get_option_quotes(self, contracts):
pass

Then implement:

MassiveProvider
EodhdProvider
IntrinioProvider
AllTickProvider
HKEXProvider

The calculation engine should not care which provider produced the data.


Example stock candle model:

market
exchange
symbol
timestamp
interval
open
high
low
close
adjusted_close
volume
vwap
provider

Example option contract model:

market
underlying
contract_symbol
expiry
strike
option_type
bid
ask
last
volume
open_interest
implied_volatility
delta
gamma
theta
vega
rho
timestamp
provider

Possible database:

PostgreSQL + TimescaleDB

Tables:

symbols
stock_candles
corporate_actions
dividends
stock_splits
option_contracts
option_quotes
option_trades
option_greeks
earnings
fundamentals
provider_metadata

For very large historical data:

Parquet

is often more efficient.

Recommended approach:

PostgreSQL / TimescaleDB
→ metadata
→ current data
→ indexed research queries
Parquet
→ large historical options
→ historical stock candles
→ bulk backtests

flowchart LR
A[Provider API / Flat File] --> B[Raw Storage]
B --> C[Validation]
C --> D[Normalization]
D --> E1[PostgreSQL]
D --> E2[Parquet]
E1 --> F[Research]
E2 --> G[Backtesting]

Raw data should ideally be retained before transformation.

This allows future corrections if:

  • provider schema changes
  • normalization logic changes
  • adjusted price logic changes
  • corporate action handling changes

Goal:

Build without paid data.

Tasks:

  • define interfaces
  • define database schema
  • implement provider adapter
  • implement data normalization
  • create mock data
  • connect Massive Free
  • connect EODHD Free

Cost:

US$0


Implement:

  • symbols
  • OHLCV
  • corporate actions
  • dividends
  • splits
  • adjusted prices
  • delisted-symbol handling

Providers:

  • EODHD
  • Massive

Implement:

  • option chain
  • expiry
  • strike
  • call / put
  • bid / ask
  • IV
  • Greeks
  • volume
  • open interest

Provider:

Massive


Implement common strategies:

Long Call
Long Put
Short Call
Short Put
Covered Call
Cash-Secured Put
Bull Call Spread
Bear Call Spread
Bull Put Spread
Bear Put Spread
Long Straddle
Long Strangle
Iron Condor
Iron Butterfly
Calendar Spread
Diagonal Spread
Box Spread

Backtest should support:

Entry Date
Entry Time
Expiry
Strike Selection
Delta Selection
Premium
Bid / Ask
Slippage
Commission
Assignment
Exercise
Early Exit
Profit Target
Stop Loss
Rolling
Expiration

Upgrade to a deeper options dataset.

Possible choices:

  • Massive Developer
  • Intrinio
  • FirstRate Data
  • Databento

Add:

  • HK symbols
  • HK OHLCV
  • dividends
  • splits
  • corporate actions
  • HK market calendars

Providers to evaluate:

  • EODHD
  • Twelve Data
  • AllTick
  • HKEX

Only after validating demand.

Tasks:

  1. identify exact instruments
  2. identify historical-depth requirements
  3. identify real-time requirements
  4. obtain vendor quotes
  5. check HKEX licensing
  6. confirm redistribution rights
  7. estimate per-user costs

PhaseMain ProvidersApproximate Monthly Cost
PrototypeMassive Free + EODHD Free$0
DevelopmentMassive Starter + EODHD~$50–60
BacktestingMassive Developer + EODHD~$100
Advanced ResearchMassive Advanced / Intrinio~$200–350+
CommercialBusiness plansQuote required
HK CommercialHK vendor + HKEX rightsPotentially high

Best balance of:

  • price
  • API quality
  • options coverage
  • historical data
  • developer experience

Best for:

  • professional analytics
  • commercial projects
  • higher-quality options research

Useful for:

  • research
  • longer historical experiments

Useful if consolidating multiple markets into fewer providers.


Best overall balance for:

  • global coverage
  • historical data
  • corporate actions
  • fundamentals

Good global API.

Interesting for Hong Kong real-time data.

Licensing requires further due diligence.


Recommended shortlist:

Good API + flat-file ecosystem.

High-quality professional option datasets.

Very attractive for offline historical backtesting.

Best suited to advanced quantitative / tick-level analysis.


For a commercial product, consider:

  • Massive Business
  • Intrinio Commercial
  • Databento
  • authorised exchange vendors

Avoid assuming that personal or individual plans permit commercial display.


Recommended approach:

HKEX
or
HKEX-authorised vendor

Do not rely on a cheap unofficial API until licensing and data provenance are confirmed.


Before signing up, send the provider these questions.

1. Do you provide US equities?
2. Do you provide Hong Kong equities?
3. Do you provide US equity options?
4. Do you provide Hong Kong stock options?
5. How many years of history are available?
6. Are delisted symbols included?
7. Are corporate actions included?

8. Is full historical option chain data available?
9. Are historical bid/ask quotes included?
10. Is open interest historical or current only?
11. Are Greeks historical or calculated only for current chains?
12. Is implied volatility historical?
13. Are trades available?
14. Are quotes available?
15. Is NBBO available?

16. Is the plan personal use only?
17. Can multiple internal users access the data?
18. Can data be displayed in a private web application?
19. Can data be displayed to paying customers?
20. Is redistribution allowed?
21. Are derived analytics permitted?
22. Are exchange fees included?
23. Are per-user exchange fees required?
24. Can data be stored permanently?
25. Can historical data remain stored after cancellation?

Subject: Market Data Licensing and Historical Options Questions
Hi,
We are evaluating your market data service for a software platform that performs
options strategy analytics and historical backtesting.
Could you please confirm the following:
1. Whether historical US equity option chains are available.
2. Whether historical bid, ask, volume, open interest, IV and Greeks are included.
3. The maximum historical depth available.
4. Whether downloaded historical data may be stored permanently for internal research.
5. Whether the subscription permits use by multiple internal users.
6. Whether derived analytics may be displayed to external customers.
7. Whether raw or delayed market data may be displayed to external customers.
8. Whether separate exchange licences or redistribution fees are required.
9. Whether you provide Hong Kong stocks or Hong Kong stock options.
10. Whether commercial pricing differs from individual pricing.
Thank you.

For the current project, the recommended path is:

Massive FREE
+
EODHD FREE

Do not subscribe to expensive feeds yet.

First build:

Data Adapter
↓
Normalized Schema
↓
Calculation Engine
↓
Strategy Engine
↓
Backtest Engine
↓
Output

When historical option backtesting is working:

Massive Options Developer

Hong Kong Stocks

using:

  • EODHD
  • Twelve Data
  • AllTick

after checking licensing.


Only add:

Hong Kong Options

after:

  • validating product demand
  • obtaining vendor quotations
  • confirming HKEX licensing requirements
  • confirming redistribution rights

The project should always maintain:

DATA
↓
CALCULATION
↓
OUTPUT

as separate layers.

This gives several benefits:

  • data providers can be replaced
  • licensing is easier to control
  • backtesting is independent
  • live trading is independent
  • commercial output can be separated from raw market data
  • the platform can start with US data and later add HK data

market-platform/
│
├── apps/
│ ├── api/
│ ├── web/
│ └── worker/
│
├── packages/
│ ├── market-data/
│ │ ├── interfaces/
│ │ ├── massive/
│ │ ├── eodhd/
│ │ ├── intrinio/
│ │ └── hk/
│ │
│ ├── calculations/
│ │ ├── pricing/
│ │ ├── greeks/
│ │ └── risk/
│ │
│ ├── strategies/
│ │ ├── vertical-spreads/
│ │ ├── condors/
│ │ ├── calendars/
│ │ └── boxes/
│ │
│ ├── backtest/
│ │ ├── engine/
│ │ ├── execution/
│ │ ├── slippage/
│ │ └── reporting/
│ │
│ └── storage/
│ ├── postgres/
│ └── parquet/
│
├── data/
│ ├── raw/
│ ├── normalized/
│ └── backtest/
│
└── docs/
├── MARKET_DATA_RESEARCH.md
├── ARCHITECTURE.md
└── LICENSING.md

https://massive.com/

https://massive.com/pricing


https://eodhd.com/


https://intrinio.com/

https://intrinio.com/pricing


https://www.alphavantage.co/

https://www.alphavantage.co/documentation/


https://twelvedata.com/

https://twelvedata.com/pricing


https://financialmodelingprep.com/


https://www.tiingo.com/


https://alltick.co/

https://alltick.co/pricing


https://databento.com/


https://firstratedata.com/


https://www.interactivebrokers.com/


https://www.hkex.com.hk/

HKEX Market Data:

https://www.hkex.com.hk/Services/Market-Data-Services


RequirementRecommended Provider
Free prototypeMassive + EODHD
US stock candlesMassive / EODHD
US dividends / splitsEODHD / Massive
Global stocksEODHD
HK stocksEODHD / Twelve Data / AllTick
US option chainMassive
US Greeks / IVMassive
Historical US optionsMassive / Intrinio / FirstRate
Tick-level US optionsDatabento / Intrinio
Broker executionIBKR
HK real-time stocksAllTick / HKEX vendor
HK optionsHKEX / authorised vendor
Commercial distributionBusiness / exchange licence

The most important conclusion from this research is:

Build the system around US market data first, keep the data layer provider-neutral, and treat Hong Kong options as a later licensing project rather than an initial programming problem.

The preferred initial stack is:

Massive
+
EODHD
+
PostgreSQL / TimescaleDB
+
Parquet

with the architecture:

Market Data
↓
Provider Adapter
↓
Normalized Storage
↓
Calculation Engine
↓
Strategy Engine
↓
Backtest Engine
↓
Result API
↓
UI / Reports / Broker

This keeps development cost low while leaving the system flexible enough to add:

  • more US data providers
  • Hong Kong equities
  • Hong Kong options
  • commercial licensing
  • real-time execution

later without redesigning the core application.