# Kairos vs Parity: one terminal charges per fill, the other has a fee written down and switched off

Both route orders to Kalshi and Polymarket from one screen. Kairos charges 0.25 cents a contract; Parity's own terms price double that if its fee comes on.

*https://predictionmarkets.tools/compare/cross-venue-trading-terminals*

- https://predictionmarkets.tools/tools/kairos.md
- https://predictionmarkets.tools/tools/predictparity.md

## Why these two, and not the other two-venue screens

Four cards in this category put Kalshi and Polymarket on one screen, and only two of them send an
order to both. [Polyrama](https://predictionmarkets.tools/tools/polyrama) reads Kalshi and trades only Polymarket; its docs call
live Kalshi orders a planned addition. [Synthesis](https://predictionmarkets.tools/tools/synthesis) lists Kalshi markets, but its
terms describe trading as wallet-signed messages to a blockchain, and whether an order is ever
routed to Kalshi is unconfirmed on its card. A reader who wants to take positions on both venues
from one ticket is choosing between [Kairos](https://predictionmarkets.tools/tools/kairos) and [Parity](https://predictionmarkets.tools/tools/predictparity).

Each adds a third venue, and in both cases the third venue does less than it appears to. Kairos's
is Predict.fun, which has no card here and whose terms this catalogue has not read. Parity's is
[Polymarket US](https://predictionmarkets.tools/tools/polymarket-us) — which serves only US residents, through a product whose own
terms bar US persons. Read both documents together and Parity's third venue is open to nobody that
both of them admit. For a reader outside the United States, these are two Kalshi-and-Polymarket
terminals, and the comparison is about what each costs and what each does with your keys.

## The fee: charged on one, written down on the other

**Kairos charges per taker fill, on a published ladder.** The entry tier takes 50 basis points of
notional at a 50-cent price, tapering to 10 at either end; makers pay nothing on every tier. That is
a quarter of a cent per contract at the midpoint, which is why the card's fee field reads 0.25.
Volume moves it: 40 basis points from 50,000 USD of 30-day volume, 30 from 99,900, 25 from one
million.

**Parity charges nothing, and says "for the time being".** Its fee schedule, dated 14 August 2026,
says there is no Parity fee on trades, settlement, membership, deposits or withdrawals. Its terms of
use, dated three months earlier and still current, describe what the fee would be: a builder fee of
1% of notional on filled taker orders and 0.5% on maker orders, peaking at a 50-cent price and
reduced towards zero near the ends. The client code already carries the builder-fee configuration,
and the referral programme already pays 30%, 3% and 2% of a net fee that is not being collected.

Put the two cards' numbers side by side at 50 cents, which neither card does:

- **Taker.** Kairos 0.25 cents a contract. Parity 0 today, and 0.5 cents under its written fee —
  double Kairos's entry tier and four times its top tier.
- **Maker.** Kairos 0 on every tier. Parity 0 today, and 0.25 cents under its written fee.

On 1,000 contracts taken at 50 cents that is 2.50 USD on Kairos, nothing on Parity today, and 5.00
USD on Parity the day the fee is switched on. So the cheaper terminal is decided by a document date,
not by a price list: Parity wins by Kairos's whole fee while its schedule says zero, and loses on
both sides of the book the day it does not.

**The venue's fee is the bigger number on either terminal, and it is the same on both.** Kalshi
charges a taker 1.75 cents a contract at 50 cents; Kairos's quarter cent is about 14% on top of
that. Polymarket's own taker coefficient runs from 0 to 0.07 by category, so a sports fill at 50
cents costs 1.25 cents a contract before either terminal adds anything.
[The fee calculator](https://predictionmarkets.tools/calculators/prediction-market-fees) does the venue half at your own price.

**The cards disagree on Kalshi's maker side, and the Kairos card is the precise one.** Parity's card
says Kalshi charges only when an order crosses the book, so a resting order is free there. Kairos's
card notes that a small set of Kalshi series carry a maker fee, and the Kalshi venue card puts it at
160 of 11,078 series in September 2026. A resting order on Kalshi is free on most series, through
either terminal, and not on all of them.

## Who signs the order

This is the part a fee comparison leaves out, and it is closer between the two than the fee is.

**On Kalshi, both hold a credential that can trade your account.** Parity asks for the API key ID
and the entire RSA private key file, keeps an encrypted copy, and says disconnecting in Parity does
not revoke the key at the exchange — you delete it at Kalshi as well. Kairos's conditional-order
engine has you store Kalshi API credentials with it for the same reason: a stop that fires while you
are asleep has to be signed by something that is not you. Parity's app says it cannot move money in
or out of Kalshi; neither card says what Kairos can do with the credential beyond trading.

**On Polymarket, the two describe themselves inconsistently in different ways.** Kairos's terms say
it holds no private keys; its API docs call the default order path "the custodial lane", signed by a
managed key in a cloud HSM, and list Kairos-custodied wallets. Only an allow-listed external-signing
lane has you sign with your own key. Parity provisions an embedded wallet through Privy at signup, or
takes an imported Polymarket private key that its docs say is uploaded to Privy; its privacy policy
meanwhile says it stores no private keys, beside a linking guide that describes Parity's copy of
your Kalshi key.

Neither document set is clean, and the honest summary is the same for both: a hosted service can
place orders in your accounts, and its legal pages and its docs do not agree on how.
[Where your key lives](https://predictionmarkets.tools/guides/where-your-key-lives) sets out what each arrangement lets a copy do.
If that is the dealbreaker, the code route is [PMXT](https://predictionmarkets.tools/tools/pmxt), one interface over several venues
running on your own keys, at the cost of writing the screen yourself.

## What "one book" means on each

Kairos shows the strongest version of the idea: one ladder per question, with a composite best bid
and offer across venues, built from pairs its matcher scores at a similarity of 0.82 or more. Parity
shows three venues' depth in columns of one ladder, filled with markets the server groups together
or, failing that, with markets whose short title is the same string. One is a model and one is a
string comparison, and both are claims about how a question is worded. Parity's own front page
showed the case: the October Fed pair it put in one ladder carried end dates of 28 and 29 October
2026.

A merged ladder is the place a terminal invites you to treat two contracts as one. The payout
follows each venue's resolution source, close time and fallback, and
[matching a question across venues](https://predictionmarkets.tools/how-to/match-a-question-across-venues) is the check neither
screen does for you.

The portfolio view runs into the same wall from the other side. Kairos's docs say there is no grand
total across venues, because Kalshi dollars and Polymarket's on-chain collateral sit in different
custody domains. Parity's Portfolio merges positions and a profit breakdown across linked venues,
while its docs say the positions settle separately and the balances are not interchangeable. The
first refuses to add two currencies; the second adds them and tells you in the docs.

## What else separates them

**Code.** Kairos publishes 102 endpoints with OpenAPI specs, a protobuf market-data stream, scoped
execution keys and an anonymous market-data tier it calls a courtesy that can be switched off.
Parity publishes nothing a developer can use, and its terms license its data for internal,
non-commercial use.

**Automation.** Kairos's Krisis engine runs stops, trailing stops, OCO and OTO brackets, TWAP and
custom conditions, with a backtest endpoint — and its docs say paper execution has been removed, and
an armed strategy with no funds attached returns success and then skips silently. Parity's stop,
take-profit, TWAP, iceberg and "Market Eye" orders run server-side, are labelled a beta that may not
be enabled on your account, fire stops as market sells off the best bid, and do not cap a sell TWAP
to the position you hold.

**What else is on the screen.** Parity carries Polymarket wallet analytics and a per-outcome
"Market Strength" share computed over wallets with more than 10,000 USD of profit; it sees
Polymarket wallets only. Kairos carries a news panel and a leaderboard-driven wallet-mirroring
feature for Polymarket and Predict.fun. Neither changes the order you place on Kalshi.

## Which to take

**Outside the United States, trading by hand, mostly with resting orders: take Parity while its fee
schedule says zero.** It is cheaper by exactly Kairos's fee, and on resting orders the gap is
nothing either way until Parity's maker fee exists. Re-read the fee schedule before any month you
expect to trade size; the terms already say what it changes to, and on that day Parity costs twice
Kairos's taker fee and charges makers where Kairos does not.

**Switch to Kairos the moment any one of four things is true:** you want to trade from code; you
need conditional orders that are live rather than rolling out; Parity's schedule stops saying zero;
or you are in the United States. On the last, be clear what you get: Kalshi and nothing else, plus
Krisis, for a quarter of a cent a taker contract on top of Kalshi's 1.75. If you do not need the
conditional orders, [Kalshi](https://predictionmarkets.tools/tools/kalshi) direct is the same contract without the quarter cent.

**Take neither if what you need is Kalshi and Polymarket US on one screen from inside the United
States.** No terminal in this catalogue offers it under its own terms. The two venues' own
apps, or a pair of API clients you run yourself, are the route that exists —
[the SDKs category](https://predictionmarkets.tools/categories/sdks-and-client-libraries) has both venues' official packages.

## FAQ

### Which of the two is cheaper?

Parity today, by exactly Kairos's fee. Parity's fee schedule of 14 August 2026 says it charges nothing of its own for the time being. Kairos charges a taker 50 basis points of notional on its entry tier, a quarter of a cent per contract at 50 cents, and makers nothing. If Parity switches on the builder fee its terms describe, it becomes the dearer of the two on both sides of the book.

### Can I use either from the United States?

Kairos, for Kalshi only. Its live restriction list blocks the US from Polymarket and Predict.fun routing and from nothing else. Parity's terms bar US persons from using the service to participate in prediction markets at all, even though it links Kalshi and Polymarket US accounts. Neither puts Kalshi and Polymarket US on one screen for a US reader under its own terms.

### Does either hold a key that can trade my Kalshi account?

Both. Parity keeps an encrypted copy of the whole Kalshi RSA private key you paste in, and Kairos's conditional-order engine has you store Kalshi API credentials with it. Disconnecting in Parity does not revoke the key at Kalshi; you delete it there too. On the Polymarket side Kairos's docs describe a default custodial signing lane while its terms say it holds no keys.

### Does either one prove two venues' markets are the same question?

No. Kairos merges books where its matcher scores two questions at a similarity of 0.82 or more; Parity pairs by server-side groups or by an identical short title. Both are claims about wording. Parity's own front-page Fed pair carried end dates of 28 and 29 October 2026, and the payout follows each venue's rules, not the merged ladder.

### Which one can I drive from code?

Kairos. It publishes 102 endpoints with OpenAPI specs, a protobuf market-data WebSocket, scoped execution keys and an anonymous market-data tier. Parity publishes no API, SDK or WebSocket, and its terms license its data for internal, non-commercial use only.