# AI Center pricing

The Coralogix AI Center pricing model is built around Coralogix Units (CUs) — the same units used across all Coralogix products.

Sending AI traces to Coralogix is billed like any other trace ingestion — there's no additional AI Center cost on top. Additional cost applies only when you enable evaluations or guardrails on your AI applications. For these policies, the calculation differs from other pillars: instead of being based on data volume (GB), it is based on the number of enabled policies and the tokens processed by your applications.

## Benefits

- **Transparency** — You can see how your usage affects cost.
- **Scalability** — Pricing scales with usage, keeping costs aligned with the value delivered.
- **Predictability** — With a fixed rate per Coralogix Unit, you can estimate monthly or project costs.

This approach keeps pricing competitive and aligned with resource consumption, helping you plan and track spend with confidence.

## Calculating Coralogix Unit usage

### Understand tokens

Every piece of text processed by an AI agent is measured in tokens. A token represents a word.

### Define policies

Coralogix applies policies to the tokens you ingest. Policies fall into two types:

- [Guardrails](https://coralogix.com/docs/user-guides/ai/guardrails/index.md)
- [Evaluations](https://coralogix.com/docs/user-guides/ai/evaluations/index.md)

### Calculate usage

To calculate usage, multiply the number of tokens by the number of enabled policies, then divide by 1,000,000.

```text
Eval usage (CUs) = (Tokens × Enabled Evaluators) / 1,000,000
Guardrails usage (CUs) = (Tokens × Enabled Guardrails) / 1,000,000
Total usage (CUs) = Eval usage + Guardrails usage
```

### Example

If an AI agent processes 10,000 tokens and uses 5 evaluators, the eval consumption is:

```text
(10,000 tokens × 5 evaluators) / 1,000,000 = 0.05 Coralogix Units
```

If that same AI agent also uses 2 guardrails, the guardrails consumption is:

```text
(10,000 tokens × 2 guardrails) / 1,000,000 = 0.02 Coralogix Units
```

Total consumption is:

```text
0.05 + 0.02 = 0.07 Coralogix Units
```

## Viewing usage in the AI Tokens usage screen

The AI Tokens usage screen breaks down Coralogix Unit consumption by policy type:

- **Evals** — Shows usage for enabled evaluators.
- **Guardrails** — Shows usage for enabled guardrails.

In graphs and visualizations, Coralogix displays the policy name with its type, such as:

- Toxicity (Eval)
- Prompt Injection (Guardrail)

## Next steps

Review the roles and permissions required to access and manage AI Center with [AI Center permissions](https://coralogix.com/docs/user-guides/ai/permissions/index.md).