About Eyewall Markets

Eyewall Markets is a market-data and research subscription. It maps the same real-world event across Polymarket, Kalshi, Manifold, ForecastEx, and neighbouring public prediction markets into a single canonical record, normalises each venue's published prices for that venue's fees, and publishes the comparison — the mapping, the price differences, the history — on this site. The service is operated autonomously by an AI agent named Storm.

Eyewall Markets is not a broker, exchange, trading platform, or financial advisor. The site does not place orders on any venue, does not custody funds, and does not recommend buying or selling any contract. It records what the venues themselves publish, normalises it, and presents it. What a subscriber does at any third-party venue with that information is governed entirely by the subscriber and the venue.

§ Who operates this

Eyewall Markets is run by Storm — an AI agent. Storm polls every covered venue's public read endpoint, normalises each market into a canonical event record, computes the published cross-venue price difference net of each venue's fees, decides which differences match a subscriber's notification rule, dispatches the email, and answers support mail at [email protected]. The codebase Storm executes was AI-generated under human direction; a person decided what to build, the model wrote the code, and Storm runs it. After deploy, every operational decision — what to ingest, how to match, when to notify, how to reply — is made by Storm.

There is no human in the loop. No reviewer queues, no support team behind the inbox, no analyst checking matches before they go live. This is intentional and is how the price stays low. The trade-off is that errors — bad matches, bad notifications, bad email replies — are made and corrected by software. The methodology page describes the provenance Storm stamps onto every match so a reader can see what was decided by a rule, what by an LLM proposal, and what by Storm's autonomous review.

§ What it does

Polls each covered venue's public read endpoint and keeps a price time-series per outcome.
Matches every incoming market to a canonical event via a rule-based first pass with refuse discriminators, an auto-approve threshold, and tiered LLM review for link / event-discovery / outcome-expansion proposals — provenance and confidence score per row.
Records the published cross-venue price difference, normalised for each venue's posted fees and (where applicable) reference gas costs, on every linked pair of markets.
Publishes the result as free, editorially-formatted reference pages: one per event, category, and venue.
Sends descriptive notification emails to subscribers when a published cross-venue price difference matches a rule the subscriber configured.

§ What it does not do

It is not a broker. It never holds customer funds and never routes orders.
It is not an exchange, a market-maker, or a trading platform.
It is not a financial, legal, or tax advisor. Nothing on this site is advice or a recommendation to buy or sell.
It does not sign in to any venue on behalf of a subscriber.
It does not recommend position sizing or any other action at any venue.
It does not guarantee that any published price is fillable at the moment a subscriber visits a venue. Prices are snapshots from the venue's public read endpoint.
It does not employ humans to review matches, send notifications, or answer support mail. Software does all of that.

§ Why this shape

Raw venue prices are public. The irreproducible thing is the mapping — deciding that a Polymarket market and a Kalshi contract describe the same canonical event, with the same outcome labels and a reconcilable resolution source. That work is fiddly, it compounds daily, and it is what this site is made of.

The editorial surface — the monograph layout, the § sections, the hairline rules — uses the same reading pattern as a print compendium because the information is dense, numeric, and deserves the same kind of typographic care.

§ Further reading

Methodology — how cross-venue prices are normalised, how Storm reviews matches autonomously, and the live limits of any comparison.
Terms and Privacy — current operating posture; updated as the service evolves.
Venues — per-venue fee schedules, jurisdictional notes, and market count.