---
title: Pricing and limits
description: One rate for every shape, and one charge per completed request.
icon: credit-card
---

**One endpoint, one rate.** Every `POST /v1/systemone` shape — Decision only, Jev only, or Jev + Decision — costs the same. Console and API calls use the same prices and credit balance.

| Work | Input per 1M tokens | Output per 1M tokens |
|---|---|---|
| Native Decision execution, Jev questions, or Jev + Decision | **$0.042** | **Free** |

Prices are in USD. One million tokens is the billing unit, not one million decisions.

There is exactly one charge per completed request, including Jev when used. The endpoint never generates a model or falls back to AI interpretation: use matching facts, or Jev questions whose answers establish the model’s inputs.

## How tokens are counted

| Shape | Billable token counts |
|---|---|
| Decision only | Local o200k counts of the submitted request and compact answer/trace response. |
| Jev only | Jev’s exact reported input and output counts, without a local recount. |
| Jev + Decision | Jev’s reported counts plus local o200k counts of the native definition and native answers/trace. Shared state, questions and Jev results are counted once. |

Usage and accounting metadata are excluded from execution output counting. Underlying provider work is not a second charge. Jev usage is not an estimate based on the length of the wire JSON.

## Two recorded calls with the same refund model

The [quickstart](/docs/quickstart) includes these October 2026 local-development measurements:

| Request | Input tokens | Output tokens | Calculated charge |
|---|---|---|---|
| Matching facts → Decision only | 652 | 273, free | $0.000027384 |
| Customer message + account facts → Jev + Decision | 1,057 | 331, free | $0.000044394 |

Each charge is `input_tokens × $0.042 ÷ 1,000,000`. The combined call’s Jev work is already included in its single total.

These costs are calculated from recorded usage and the published rate, not fixed per-decision prices. Actual cost depends on the request and response token counts; formatting can change those counts.

## Read usage and charges

Public execution returns `answers`, `receipt` and `usage: {input_tokens, output_tokens}`. It does not include timing, billing, provider metadata or balance. The console shows settled charges, balance and separate provider counts. Native billing uses API counts; Jev-only billing uses provider counts; combined billing uses both without double counting.

The console may display tiny nonzero charges as **less than $0.001**. That display does not make the request free.

The console’s manual **Check** is free. It is a console operation; there is no public validation or model-authoring call. **Cancel** aborts the browser request but does not guarantee backend work stops or that no charge is incurred.

## Completed Jev with a model failure

If Jev completes but the following native model fails, the call returns HTTP `200` with Jev’s answers, trace and `decision_error`. The completed Jev operation is charged once at the single rate. Check `decision_error`, not only the HTTP status. See [Errors](/docs/resources/errors).

## Preview credits and capacity

New accounts start without credits. To get your first credits, email [support@aletyx.com](mailto:support@aletyx.com) from the address you signed up with. Jev calls start whenever available balance is positive. Completed Jev work is charged in full and may leave a negative balance; subsequent calls stop until credits are added.

Decision-only execution reserves funds before work, settles the completed charge, and releases unused reservations. A positive balance can still be too small for its reservation, returning `402 insufficient_balance`. Failed operations release reservations; completed Jev work followed by a model failure is still billable.

The account ledger tracks credits and charges separately from the models, states and receipts your application keeps. Account and balance endpoints remain console/internal only.

The preview does not publish a universal throughput allowance. For capacity or volume needs, contact [support@aletyx.com](mailto:support@aletyx.com). Sign up in the [console](https://console.aityx.ai) to create an API key.
