Institutional datasheet

This is what dealer positioning looks like, computed.

FlashAlpha computes the dealer-positioning, flow, volatility and variance-premium layer of the US options market and serves it as production infrastructure: the same metric definitions live, historical, and streamed. Research teams test a hypothesis the same day instead of spending two quarters building the pipeline that would let them.

Datasheet updated 2026-07-28. Every figure below is documented in the linked methodology and due-diligence pages, including the stated limitations.

Net gamma exposure by strike schematic
positive gamma negative gamma GEX / DEX / VEX / CHEX, per strike, live and since 2017
Specification
What it is
Computed options analytics as an API, not a terminal or a dashboard subscription.
Coverage
6,000+ US equities and ETFs, major cash indices, and CME futures (gold, crypto, rates, grains, index) with Black-76 options-on-futures.
Analytics
Gamma, delta, vanna and charm exposure (GEX/DEX/VEX/CHEX); gamma flip, call and put walls; max pain and pin probability; 0DTE regime and expected move; SVI volatility surfaces; variance risk premium; options flow with dealer-position inference; 15 BSM Greeks.
History
Point-in-time replay from 2017-01-03 at minute resolution on the flagship symbols, with response shapes identical to live via an at= selector.
Delivery
REST, commercial WebSocket streaming, MCP server for AI research agents, SDKs for Python, JavaScript, C#, Go and Java.
Licensing
Internal research and trading only. No redistribution, embedding, white-label, or publishing on any tier.
Entity
FlashAlpha LTD, Cyprus (EU). EU-hosted infrastructure.
The case

Why desks license it

Research starts on day one

Reproducing this layer in-house means an options-chain pipeline, an IV solver, a surface fitter, exposure aggregation, and years of clean history before the first hypothesis is testable. That is a multi-quarter engineering programme with permanent maintenance attached. Here it is one call.

Backtest and production are the same contract

Every core analytic has a historical twin with an identical response shape. A strategy validated on history ships to production without re-derivation, and the live and historical engines are held numerically in parity by an automated cross-engine test.

Aggregates published almost nowhere else

Aggregate vanna and charm exposure are available from very few public APIs. Desks using them as model features report several of their strongest predictors coming from this layer, precisely because the inputs are not in everyone else's feature store.

A vendor that documents its own limits

The methodology whitepaper states the dealer-positioning assumption, European pricing on American contracts, the un-versioned history, the absent halt detection, and the reconciliation blind spots. Diligence teams can evaluate the model rather than guess at it.

Delivery

One metric definition, three ways to read it

The same computation serves all three, which is why a backtest and a production signal cannot silently disagree.

Research

Historical API

Point-in-time replay since 2017 with an at= selector, on a host isolated from live serving. Integrity posture.

Production

REST API

Request and response on your schedule, with per-snapshot quality telemetry your pipeline can gate on. Reference.

Real time

WebSocket streaming

The same metrics pushed as they recompute, subscribed by symbol, metric and cadence. How it works.

Commercial tiers

All commercial tiers are licensed for the client organisation's own internal research and trading. Standard self-serve plans (Free, Basic, Growth, Alpha) cover personal and single-team use and are published at /pricing.

Professional
from$2,500/mo
Dedicated infrastructure

A dedicated node with reserved throughput for high-rate polling and intraday research loops. Full analytics stack, live and historical parity.

Streaming
from$4,500/mo
Real-time push

WebSocket push of 500+ computed metrics, bundling a dedicated node, streaming compute and data licence with 500 concurrent subscriptions; further capacity in blocks of 500.

Enterprise
Custom
Scoped to requirements

Multi-region and reserved capacity, custom metrics and proprietary signals engineered into your endpoint, procurement and compliance scoping.

Commercial agreements support annual invoicing by bank transfer; master service agreements, data-processing terms, and SLA riders are handled during scoping. New computations requested by a client are typically delivered in days, as a maintained first-class field on their endpoint rather than a one-off extract.

Provenance you can assert on

Published documentation tells you how a number is built. It does not tell you how fresh the inputs behind this response were. Every FlashAlpha response answers that itself, per feed:

  • Live - data_as_of reports when each upstream feed last delivered to the node that answered: equity and index spot, their option chains, futures, the classified tape, settled open interest and the macro series, each separately. Spot and options are distinct because they arrive over different pipes and fail independently.
  • Historical - archive_as_of reports the vintage of the archive rows actually replayed for the timestamp requested. This is the field that makes an archive gap visible: request a moment with no row and the query returns the most recent earlier one, and nothing else in the response distinguishes the two.
  • Continuously verifiable - both are on every call and fixed in shape, so a client asserts on them in code rather than taking a freshness claim on trust.

It evidences that a feed delivered recently, not that every contract in a chain is equally current, and each feed is read against its own cadence. Field reference and methodology.

For your diligence team

The material an operational and data-diligence review normally has to request is already published, and the freshness of every answer is carried in the answer itself:

  • Methodology whitepaper - pricing model, surface construction, exposure definitions, point-in-time integrity, validation, and an explicit assumptions-and-limitations section. Versioned at a stable URL with a citation block.
  • Data Quality & Monitoring - the verification architecture: cross-source reconciliation, fail-closed serving guards, fleet-wide job health, and the no-silent-restatement policy.
  • Fleet monitoring and operational controls - cross-source reconciliation, job-health registry, fail-closed serving, and the data-availability control, including what these controls deliberately do not do.
  • Security, topology and company overview - hosting, encryption, access control, service topology, backups and recovery, support and escalation, legal entity, and current attestation status stated plainly.
  • Per-response provenance - every response carries data_as_of, reporting when each upstream feed last delivered: equity and index spot, their option chains, futures, the classified tape, settled open interest and the macro series, each separately. Historical responses add archive_as_of, the vintage of the archive rows actually replayed for the timestamp requested. Freshness is asserted per call and machine-readable, not asserted once in a document.
  • Status page - 90-day component availability measured by an off-fleet probe over the real request path.
  • API reference, Historical API, and machine-readable specification - full endpoint surface with real captured response samples.
  • Terms, Privacy, and Risk disclosure.

Questions a questionnaire will raise and this page answers directly: FlashAlpha is not currently SOC 2 or ISO 27001 certified, and says so; history is retained but not restatement-versioned; exposure metrics are a structural model of hedging pressure, not observed dealer inventory. Where a commitment is Enterprise-scoped rather than standing, it is labelled as such. data_as_of evidences that a feed delivered recently; it is not a claim that every contract in a chain is equally current, and an illiquid strike may not have quoted for hours while its feed is healthy.

Scope a deployment

Tell us the research or trading problem and we will size the tier, the symbol universe, and the throughput against it. Commercial enquiries are answered within one business day by the engineering team, not a sales queue.