IRYXby Zyrex
Meet:
Iryx by Zyrex
Say:
EYE-riks, with a Y
Product:
Iryx API v1

Iryx decides.

The decision model by Zyrex. It takes a context. It returns typed decisions, each with measured confidence and the full spectrum. It doesn't write text.

Response:~74 msmedian per request, through the API
Math:100%exact, 33 of 33
Main test:82%225 questions
Blind tests:73–86%2 tests, new domains

Measured on Zyrex's own answer key on 2026-09-23, on short texts in Portuguese. See the numbers.

Try it now

Pick a message. Iryx opens each decision as a spectrum, scores the options and keeps the most likely one. The band (low, medium or high) is the confidence cut Zyrex recommends: on low, check before you act.

Try it now · session 001REC
Input:
context + decisions
Output:
value, confidence, spectrum
Case:
real test
See the API request POST /v1/decide

A demo with ready-made answers: this page doesn't call the model. The sneaker exchange is a real test, run on 2026-09-24 with a made-up message in Portuguese; the test kept the chosen value of each decision, and the rest of the spectrum shows as a sum. The other two cases are illustrative examples. The band is computed on this page, with the recommended cuts; the API returns the value, the confidence and the spectrum.

[ benchmarks · own answer key ]

Iryx in numbers

Four measures, one answer key. Where Iryx does well and where it still misses the most.

Questions:
425
Tests:
main · 2 blind
Answer key:
own
Date:
2026-09-23

Four measures,
one answer key.

Accuracy on each test, on the same scale. The blind tests bring subjects Iryx never saw. The second one is where it still misses the most, and the chart shows it.

~74 msper request of ~3 decisions, median, measured through the API on the same machine.
100%correct on the 33 questions with a computable answer.
86% · 73%on the two blind tests, with new subjects. It still misses more on the second: there, check the runner-up.
CH 02 · 10 seconds of responsesREC
00.0 s

Each pixel is one complete request, with ~3 decisions, at the time measured through the API on the same machine. Over the internet, add the network time.

All the numbers

All numbers for Iryx
MeasureIryxN
Math, exact answer100%33
Main test82%225
Blind test 1, new domains86%100
Blind test 2, other new domains73%100
Response, median per request~74 ms200

How to read it. Text has no neutral answer key: we use Zyrex's own. The 33 math questions are part of the main test, and there the truth is absolute. The blind tests bring subjects Iryx never saw; the second is the hardest for it today. All measured on 2026-09-23, on short texts in Portuguese.

Pricing

Billed per decision, never per token. No free plan: Try it now shows the product without an account. Pay yearly and get 2 months free. Prices in Brazilian reais.

[ plus ]

Plus

R$ 110 per month

or R$ 1,100 per year, with 2 months free

Decisions per month:
3,600,000
Overage:
R$ 0.030 per 1,000
Talk to us↗

[ ultra ]

Ultra

R$ 490 per month

or R$ 4,900 per year, with 2 months free

Decisions per month:
17,000,000
Overage:
R$ 0.029 per 1,000
Talk to us↗

[ business ]

Business

R$ 1,990 per month

or R$ 19,900 per year, with 2 months free

Decisions per month:
70,000,000
Overage:
R$ 0.028 per 1,000
Talk to us↗

[ dedicated business ]

Dedicated Business

from R$ 4,900 per month

For high volume or a setup of your own. We'll work out the format together.

Talk to us↗

[ pay as you go ]

Pay as you go

R$ 0.033 per 1,000 decisions

No monthly fee. Minimum top-up of R$ 50.

Talk to us↗

Prices in Brazilian reais (R$). The unit is the decision: a request with three decisions counts as three. Each plan's overage costs that plan's price per 1,000. Business volume is confirmed when we talk.

Where we're going

A direction, not a measured result.

[ brand consistency ]

Does the piece follow the system?

Iryx checking whether a piece follows its design system (color, type, spacing), with a confidence number.

Track:
art direction
Status:
research

[ art direction at scale ]

A thousand pieces, one aesthetic.

Small, fast decisions that keep one aesthetic across thousands of pieces.

Track:
design systems
Status:
research

[ second opinion ]

The agent asks. Iryx decides.

Software agents checking with Iryx before they act, with the confidence band saying when to stop and ask a person.

Track:
agents
Status:
internal use

How Iryx was made:

How to start

The dashboard isn't live yet. For now, everything starts by email.

  1. Talk to us

    Tell us what you want to decide, and at what volume.

  2. Pick a plan

    Plus, Ultra, Business or pay as you go. We'll help you estimate your decisions per month.

  3. Call the API

    With your key, one POST /v1/decide request carries the context and the decisions. Back come the value, the confidence and the spectrum of each one.

FAQ

What is Iryx?

A decision model. It takes a context and closed questions (pick one option, yes or no, a level on a scale) and returns each answer with measured confidence and the full spectrum. It doesn't write text.

How do you say it?

EYE-riks. Spelled with a Y: I, R, Y, X.

What counts as a decision?

Each question answered. A request with three decisions counts as three. Billing is per decision, never per token.

Is there a free plan?

No. Try it now, on this page, shows the product with no account and no credit. To use it for real, plans start at R$ 110 per month, and pay as you go starts with a R$ 50 top-up.

Where does it still miss the most?

On subjects it has never seen. On blind test 2, with four new domains, it scored 73% (100 questions), against 82% on the main test. There, more decisions land in the low band: decide together, looking at the runner-up.

What happens to what I send?

Today, the API doesn't store the context or the answers: it reads, decides in memory and replies. The log for each request keeps only operational data: the time, the method, the route, the status, the request id, the number of decisions and the duration, plus the error type when something fails. The hosted service isn't live yet; its rules will be in the terms of use before the first customer.

Does it work in English?

The API takes any text. The numbers on this page were measured on short texts in Portuguese; in English, quality hasn't been measured yet.

How was it made?

Classified.