> ## Documentation Index
> Fetch the complete documentation index at: https://developers.argosidentity.com/llms.txt
> Use this file to discover all available pages before exploring further.

# How credits are charged

> Which actions deduct credits, and how to read the numbers on the Credit Usage tab.

Credit rates are **set per project by your ARGOS contact**. There is no rate configuration screen in a customer account; you review the results on the [credit usage](/en/omni/dashboard/project/credit-usage) tab.

This page explains **which action produced** each number on that tab.

## The three categories

| Category    | What it pays for                              | When it is deducted                                            |
| ----------- | --------------------------------------------- | -------------------------------------------------------------- |
| Environment | Keeping a workflow available                  | Automatically at 00:00 UTC daily, regardless of project status |
| Operation   | Handling an API request                       | When the API response succeeds                                 |
| Tools       | Calling an external engine during an analysis | Only when the tool call succeeds                               |

<Warning>
  Environment keeps billing for as long as the project exists. In every status — service, trial, and closed alike — it is deducted daily, even if you never run a workflow. Delete test projects you are not using.
</Warning>

## The actions that appear in your transactions

The **Operation** column of the transaction history records these actions.

| Category    | Action                       | Billing unit                                                                    | Base rate                                                                           |
| ----------- | ---------------------------- | ------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------- |
| Environment | Keeping a workflow available | Per day                                                                         | 1.20                                                                                |
| Operation   | Creating a profile           | Per profile created                                                             | 0.03                                                                                |
| Operation   | Adding a file item           | Per page of the file. **Adding a text item is free**                            | 0.05                                                                                |
| Operation   | Running an analysis          | Per run                                                                         | By model — [see below](#an-analysis-is-charged-by-step-count)                       |
| Tools       | Calling an engine            | Per call. Shown as the engine code (`text-similarity`, `aml-search`, and so on) | [Engine catalog](/en/omni/dashboard/workflow-setup/engine-selection#engine-catalog) |

One analysis is recorded across several rows — the analysis run itself is one Operation row, and each engine it called is one Tools row.

Tools deducts only calls with `success: true`, and tools with no rate are free. The per-call price of each engine is in the catalog under [AI model and engines](/en/omni/dashboard/workflow-setup/engine-selection#engine-catalog).

## An analysis is charged by step count

An analysis does not cost a fixed amount. The **base charge covers up to 5 steps**, and **from the 6th step onward** each step adds a surcharge.

```
final credits = base rate + (base rate × 0.2) × (step count − 5)
```

* At 5 steps or fewer the surcharge is 0, so you pay **exactly the base rate**.
* From the 6th step on, each step adds **20% of the base rate**.
* The base rate depends on the AI model used — [rate per model](/en/omni/dashboard/workflow-setup/engine-selection#the-models-you-can-choose). Your ARGOS contact adjusts it per project.

With the default model (Gemma4 12B) and its base rate of 0.18, for example:

| Steps | Calculation         | Final credits |
| ----- | ------------------- | ------------- |
| 5     | 0.18 (no surcharge) | **0.18**      |
| 10    | 0.18 + 0.036 × 5    | **0.36**      |

<Note>
  The step count equals the number of actions the playbook defines. You can count them in the **Playbook Actions** list on the creation-complete and edit screens, and the totals at the top of the Audit tab show them too. Removing an action lowers the cost of every analysis by that much.
</Note>

Check what was actually deducted in the **Final credits** column of the transaction history. To see which analysis it was, follow the `analysis_` link in the **Resource** column. `run_analysis` rows are computed with this formula, so their BASE and Quantity columns show `—`.

## Reducing your bill

* **Keep the playbook at 5 actions or fewer.** From the 6th step, every analysis carries a surcharge. Where one action can cover two checks, combine the clauses in your policy.
* **Delete projects you are not using.** Environment is billed daily even for a closed project.
* Add material that can be plain text as a **text item** instead of a file. Files are billed per page; adding text is free.
* Do not put unnecessary external lookups in your policy. One tool call is one Tools charge — [AI model and engine selection](/en/omni/dashboard/workflow-setup/engine-selection).
* Do not verify with an external engine what you could settle by comparing within the documents. Text Verifier makes no external lookup and has the lowest price.

## Related pages

* [Credit usage](/en/omni/dashboard/project/credit-usage) — reviewing usage by period and the transaction history
* [AI model and engines](/en/omni/dashboard/workflow-setup/engine-selection) — what each engine does and its per-call price
* [Project management](/en/omni/dashboard/project-management) — project status and deletion