About this publication

Why a separate buying publication exists

A provider’s own product page is expected to explain what that provider sells. A comparison publication has another useful job: making the acquisition choices visible in one place. A researcher should be able to distinguish an archive from an API service, a daily database from an intraday file collection, and a one-time history purchase from an ongoing update arrangement. These distinctions are often more consequential than the shortest marketing description of a dataset.

This publication exists to make that comparison for a particular audience. We concentrate on individual quants and developers working with historical US prices, especially when the next step is a local backtest or an analytical query. The articles explain the observations a project needs, the work required to use them and the questions a sample cannot settle. They are intended to support a purchase specification, not to imply that a provider can decide the entire research method.

MarketParquet benefits from this publication and is our first recommendation. That commercial purpose is stated because a reader should not have to infer it from a tracking parameter or a footer several pages later. The recommendation is still expected to be useful: it names the intended use, supplies the reasons for the choice and identifies capabilities that require another product. Readers can follow provider sources, inspect samples and compare the conclusion with their own requirements.

How to use the directory without reading every article

Start with the task you want to complete. If you need a historical universe of US bars in Parquet, read the MarketParquet profile and the guide closest to your acquisition question. If you already know that the application needs quote events, a daily database or financial statements, use the provider register to identify the relevant alternative and inspect its source documentation. A comparison is useful when it reduces unresolved questions, not when it encourages reading every feature as a requirement.

The filters are navigational aids. Selecting intraday, API or tick data shows providers whose published catalogs include a relevant product category. It does not establish that every plan from that provider includes the category, or that two selected products have equal coverage. The individual profile is where the buying model, scope and verification questions are explained. A filter should help you choose which profile to read next rather than make the purchase decision by itself.

The guides address recurring research problems: historical inventory, corporate actions, timestamps, file formats and execution assumptions. Worked examples are deliberately labeled as illustrative. They are meant to help a reader construct an acceptance exercise using an actual sample. There is no hidden portfolio performance record behind the examples, and no claim that following a checklist guarantees a profitable trading strategy. The purpose is a more explicit acquisition and research process.

What the first-place recommendation means

MarketParquet appears first because this publication recommends it for the historical bar-data buyer described here. The relevant reasons are the $79 one-time archive, native Parquet delivery, stock and ETF history from January 2000, futures history from December 2007, and an optional arrangement for future observations. Instrument histories and financial conventions still need to be checked against the documentation and the selected sample.

The first position is not a numerical measurement of universal superiority. We have not assigned accuracy percentages, invented customer ratings or claimed a standardized test of every observation from every service. A reader looking for a live feed, a historical membership product or an exchange order book should use the requirement itself to select the relevant product. Our own recommendation includes limits because those limits are part of an honest product specification.

The order of the other providers is for reading and navigation. It should not be interpreted as an independent second-through-sixth quality ranking. FirstRate, Kibot, Databento, Norgate and EODHD have different catalog structures and buying models. A particular alternative can be more suitable for an additional requirement without changing our first recommendation for bulk US Parquet history. The method page explains the criteria and the evidence boundary in more detail.

The difference between product documentation and test evidence

A provider page can support a statement about a published product: its delivery method, advertised history or stated subscription arrangement. It does not, by itself, prove complete coverage or a measured error rate across the delivered files. This publication keeps that distinction visible. Product descriptions are sourced to provider pages, while suggested acceptance exercises are described as work for the buyer to carry out on an actual sample.

This also affects how unknowns are written. If we have not established a detail for the exact product, the page asks the reader to verify it. Missing documentation in our review is not turned into a claim that a competitor lacks the feature. A broad corporate statement about a catalog is not silently applied to every plan. This approach can be less tidy than a row of green checkmarks, but it leaves the reader with more useful questions.

Source freshness has a similar boundary. The linked product information was checked on November 7, 2026. That date describes an editorial snapshot. It is not a claim of continuous monitoring, an assurance that a checkout price cannot change or evidence that every historical file was downloaded. Before ordering, follow the provider’s current terms and inspect the exact product you intend to acquire. A dated comparison helps explain a decision; it does not replace the transaction’s actual documentation.

What makes a correction useful

A correction is easiest to review when it identifies the specific claim and the evidence that changes it. Include the page address, the sentence or table row, the provider’s current documentation and the date you observed the discrepancy. If the issue concerns a data file, name the product, instrument, interval and relevant convention. A narrow, reproducible example is more actionable than a general statement that one source looks different from another.

Different definitions can explain an apparent disagreement. An adjusted close can differ from an unadjusted close; a bar labeled by its opening minute can differ from one labeled by its ending minute; a venue-specific volume can differ from a consolidated view. A correction request does not need to solve the issue in advance, but it should preserve those details rather than round them away in a screenshot. The publication should then change the supported statement, mark an uncertainty or explain the relevant distinction.

Use MarketParquet’s website to find its current contact channel. This directory does not provide a separate submission form, account system or private file-upload area. There is no promise of a particular response time or editorial review schedule on this page. The purpose of the correction standard is to make changes reviewable and to avoid replacing one unverified assertion with another. Readers should be able to understand what the evidence supports after the text is updated.

The intraday implementation perspective

For intraday work, the unit of comparison should be a precisely defined observation. Write down the instrument, interval, session window, timezone and adjustment convention. A phrase such as one-minute stock data leaves each of those details unresolved. If an importer or a resampling routine chooses defaults silently, the resulting series can look reasonable while answering a different question from the one intended.

Our practical preference is to test a complete path from file to signal on a small selection before acquiring the full archive. Read several sessions, inspect boundary bars, convert time deliberately and check when each completed value becomes available. Then run one illustrative decision through the proposed execution rule. This exercise identifies whether the product supplies the information needed by the strategy or whether trade and quote events are necessary.

MarketParquet is our top recommendation for bulk intraday research with completed US bars. Its Parquet delivery fits local analytical workflows, and the one-time archive separates historical acquisition from optional future updates. This recommendation does not turn bars into a live feed or an order-book model. Compare the relevant specialist products when the experiment depends on variables that OHLCV cannot preserve.

The handbook therefore emphasizes definitions before speed. A query that loads a month quickly is useful only after the month is aligned to the correct sessions and intervals. Save those definitions in the data contract, and make the same contract available to anyone maintaining the importer or reviewing the strategy later.