Options Data & Flow Analytics

Options data is sold to two different people, and most of this listing sorts itself by which one you are. If you write code and want the chain, the greeks or the tape as data, the question is how far back, how fine-grained and how cleaned — ThetaData for every OPRA tick from $40 a month, ORATS for a surface somebody already smoothed, and Cboe DataShop for history bought by the file. If you want to look at the options market rather than download it, the products here are browsers over the same OPRA feed, and most of them hand you no data at all.

• Every OPRA quote and trade, cheapest — ThetaData from $40 a month, free end-of-day.
• Smoothed IV surfaces and greeks over REST — ORATS from $199, or IVolatility from $79 with history to 2005.
• History as files, straight from the exchange — Cboe DataShop, from $50 a trading day.
• The options tape and dealer exposure with an API — Unusual Whales from $150, Quant Data from $149.99.
• Dealer positioning rebuilt trade by trade — Volland, per-strike exposure charts from $150.
• Watching the market from a screen, with the option tape back to 2011 — LiveVol from $105.
• Earnings moves and IV rankings, no API — Market Chameleon, $99.
• Laying out one position before you place it — OptionStrat or Option Samurai.
• Running spreads as no-code bots through your own broker — Option Alpha, $149 a month, or free with a funded partner-broker account.
• Replaying a position through past days, five minutes at a time — OptionNET Explorer, Windows, 535 GBP a year.

Four jobs, and they do not compete with each other

Options data is the chain and the tape as data: NBBO bid and ask per contract, trades, open interest, history. Judged on what fraction of OPRA actually reaches you — many vendors filter quotes to survive the message rate, and a quote file that is "complete" at one-minute snapshots is not complete at the tick — and on how far back each granularity goes, because end-of-day, minute and tick history almost never start on the same date.

Volatility is the chain with the modelling done: implied volatility per contract, a fitted surface, constant-maturity indexes, earnings-move statistics. What you are buying is somebody else's choices about dividends, rates and which quotes to trust, so the thing to read is how the surface is fitted, not how many fields the endpoint returns.

Options flow is the tape interpreted: prints grouped into orders, labelled, scored for sentiment, and turned into a dealer-exposure curve. This is the job where the vendor's model is most of the product and the least visible part of it.

Strategy analysis is tooling for one position — a P&L chart across price and time, a scanner that runs a structure across the market, a backtest of a rule. These products compute against a chain you never see. Nothing here suggests a trade: the structure, the strikes and the target price are the reader's, and the tool draws what they do.

A price comparison across jobs says nothing. An archive of OPRA trades and a P&L visualiser both cost about $100 a month and have nothing else in common.

What to check before you read a pricing page

Exchange fees, and who pays them. Real-time US options data is OPRA's, licensed on its own schedule, and a professional subscriber pays several times what a non-professional does. Some vendors bundle it, some itemise it: LiveVol Core is $105 a month, and a live non-professional seat came to $128.95 in the cart once the exchange fees were added. Read the line under the price.

Delayed or live. Market Chameleon is fifteen minutes delayed on every plan, and so is the free tier almost everywhere else. Delay is a licensing decision, not a technical one — it is how a vendor avoids the per-subscriber real-time fee — which is why no plan upgrade on a delayed-only product makes it live.

How the greeks were made. Per-tick Black-Scholes against the underlying's price at that moment (ThetaData), a smoothed surface reconciled with put-call parity (ORATS), or the vendor's own raw IV (IVolatility). All three are called "greeks", and they are three different calculations. Greeks has the longer version.

History by granularity, not by headline. "History to 2012" usually means one dataset at one resolution. ThetaData's options history is tick-level from 2016 on its middle tier and from 2012 only on Pro; ORATS is end-of-day from 2007 but one-minute only from 2020; Unusual Whales' $150 API looks back two years, Quant Data's a year and a little more.

API or no API. Six products in this listing have no public data API at all — LiveVol, Market Chameleon, OptionStrat, Option Samurai, OptionNET Explorer and Option Alpha — and Market Chameleon's and OptionStrat's terms forbid scraping as well. That is a fine answer for a person and the wrong one for a program, and it is the first field to read on a card.

Commercial use. Retail options plans are almost always personal-use licences. ThetaData sells commercial use as a separate $2,000 a month plan; IVolatility's business use is quote-only; on the dashboards, redistribution is negotiated by email at Market Chameleon and forbidden at OptionStrat. If anything you build will show the data to somebody else, read the licence before the tier table.

Flow, exposure and sentiment are inferences

The options tape itself is a record — every print OPRA disseminated, with time, price, size and exchange. Everything sold on top of it is a model, and the vocabulary hides that.

"Sweep", "block", "split". These are labels a vendor applies by grouping prints that arrived close together across exchanges. The grouping window and the size thresholds are the vendor's, so the same minute of tape can read as one sweep on one site and four trades on another.

"Bullish" and "bearish". A sentiment tag is a guess at whether the buyer or the seller initiated the trade, usually from where it printed against the bid and ask. A trade at the mid has no honest answer, and a leg of a spread tagged on its own has a misleading one.

Gamma exposure. Nobody outside a dealer's books knows which side the dealer took. The common shortcut takes open interest and assumes customers bought every put and sold every call; Volland argues against exactly that and classifies each trade instead. Both are estimates. A product that prints a GEX number without saying which assumption made it is selling a figure nobody can check — including the vendor.

None of this makes flow products useless. It makes them a view of the tape through a stated model, and the model is the thing to compare.

What the money looks like

A monthly tier by history and granularity. The API vendors. ThetaData runs $40, $80 and $160 for options, separated by how far back the data goes and whether trades and higher-order greeks are in it; IVolatility $79 to $299 by request count; ORATS $199 to $899, where the jump from $199 to $299 is the move from delayed to live.

Per file. Cboe DataShop prices an order: $50 for a full-market day of end-of-day summaries, capped at $300 a calendar month, $7,200 for the whole archive back to 2012, and $740 for a single day of one-minute quotes. You know the number before you spend it, and there is no free tier.

A dashboard subscription, with the API sold separately. The flow vendors split the two. Unusual Whales is $50 to $120 for the web product and $150 up for a key; Quant Data is $74.99 for the dashboard and $149.99 for the API, sold as two products rather than two tiers.

A screen, priced per seat. LiveVol Pro is $420 a month per user, and the option tape replay is what the jump from Core buys. Market Chameleon, OptionStrat and Option Samurai are $40 to $100 a month, and what changes between their tiers is live data, alerts and the number of legs — not any data you can take away.

Where this category connects

A general market data API that also sells an options plan is primary in market data APIs and listed here as a secondary member: Massive sells options as its own subscription from $29 a month, delayed, with real-time from $199, and Databento sells OPRA as one dataset among its exchange feeds from $199. Both are bought for equities as often as for options, and both hand you the raw tape with no surface and no flow layer on top. The traffic runs the other way too: ThetaData and Cboe DataShop are primary here and appear there, because both sell stock and index data on their own terms. The options data collection is the wider cut — every card that covers options at all, nearly eighty of them, including the journals that import a spread and the aggregation APIs that read one out of an account.

If you are pricing the instruments yourself rather than buying the output, QuantLib builds the curves and prices the contracts and supplies no data at all. A year of OPRA ticks has to land somewhere, which is tick data storage. And testing whether a rule built on any of this held up is a job for a backtesting framework with the history underneath it, not for a dashboard's own chart.