> For the complete documentation index, see [llms.txt](https://public-docs.anvilogic.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://public-docs.anvilogic.com/get-started/ai-operating-system-pricing.md).

# AI Operating System Pricing

## Monitor and Control AI Credit Consumption

Anvilogic prices AI usage with a single metric: **credits**. Credits are one currency across every AI surface in the platform, so there are no separate quotas per feature and no surprise line items.

Credits meter the AI and agent work the platform performs for you. They do not meter the data you send. Onboarding a new log source does not raise your credit consumption, and neither does a spike in event volume.

### Benefits

* **Credit monitoring and budgets.** See consumption by product, by day, and by workflow. Set daily budgets and adjust them whenever your operations change.
* **Full transparency.** Every run writes a usage record showing credits consumed and credits remaining. Fixed-rate activities are published rates, so you know the cost before the run.
* **Scalable and security first.** Credits are shared across the platform, so your capacity moves to whatever your team is working on that week. Approval gates, RBAC, and per-product budgets govern what agents run and how much they spend.

### How you buy

* You purchase credits in **yearly packs**. A pack is sized by an average daily consumption rate, then multiplied across the year.
* Inside the pack, you set **daily budgets** per product. Budgets are adjustable at any time, so a heavy detection engineering week or an incident surge does not require a new purchase.

#### Example credit packs

| Credits per day | Credits per year |
| --------------- | ---------------- |
| 300             | 109,500          |

Additional pack sizes are available. Your Anvilogic account team will size a pack against your alert volume and automation plans.

### How to calculate what you need

Start with alert volume. Triage is the largest and most predictable part of most teams' consumption, so it anchors the estimate.

1. **Count your average alerts per day.** Use a 30 day average, not a peak day.
2. **Double it.** The 2X pad covers the AI Chat sessions, Blueprint runs, and MCP tool calls that happen on top of triage.
3. **Round up to the nearest pack.**

#### Example

> **You average 150 alerts per day.**
>
> 150 alerts, doubled, is 300.
>
> **Start with the 300 credit per day pack. That is 109,500 credits for the year.**

That is the fastest way to get to a number. After 30 days of real usage, open the credit dashboard, compare actual daily consumption against the estimate, and adjust your daily budgets. If you plan to put Blueprints into production across several workflows in the first quarter, size above the 2X pad rather than at it.

### How credits are consumed

Two kinds of activity draw from the pack.

**Fixed rate.** A published credit cost per event. The cost does not vary with token count, tool calls, or run duration, so you know it in advance.

**Metered.** The cost scales with the work performed, because a Blueprint that onboards a data source and one that enriches an alert are not the same amount of work. Published averages give you a planning number, and budgets and per-run tracking apply to every metered activity.

<table><thead><tr><th width="149.109375">Product</th><th width="146.7109375">Activity</th><th width="120.09375">Type</th><th width="191.12890625">Description</th><th>Credits consumed</th></tr></thead><tbody><tr><td>Triage Agent</td><td>Threat Identifier verdict</td><td>Fixed rate</td><td>Verdict on a Threat Identifier alert.</td><td><strong>1 per alert</strong></td></tr><tr><td>Triage Agent</td><td>Threat Scenario verdict</td><td>Fixed rate</td><td>Verdict on a Threat Scenario alert.</td><td><strong>3 per alert</strong></td></tr><tr><td>Anvilogic MCP Server</td><td>Tool call</td><td>Fixed rate</td><td>A single tool call executed through the Anvilogic MCP Server.</td><td><strong>1 credit per tool call</strong></td></tr><tr><td>Blueprints</td><td>Workflow run</td><td>Metered</td><td>Build and run your own agentic workflow across onboarding, detection, health, and investigation.</td><td>Averages <strong>30 credits per session</strong>*.</td></tr><tr><td>AI Chat</td><td>Ad-hoc chat interaction</td><td>Metered</td><td>Ad hoc chat interaction against your data, detections, and cases.</td><td>Averages <strong>2 credits per session</strong>*.</td></tr><tr><td>Federated Search</td><td>Ad-hoc chat interaction</td><td>Metered</td><td>Ad hoc search against your data and consolidate data sets.</td><td>Averages <strong>1 credits per session</strong>*.</td></tr></tbody></table>

Anvilogic may amend the rates in this table from time to time as reasonably required.

{% hint style="info" %}
**Metered\*** - for metered activity the above table represents the average credit consumption per session across existing Anvilogic customers.  This is not guaranteed for each session.
{% endhint %}

### Budgets and usage thresholds

You set budgets **per product, per day**, so consumption in one area cannot drain the pack for the rest. A runaway Blueprint does not take your Triage Agent offline.

* Set a daily budget for Blueprints, Triage Agent, AI Chat, and the Anvilogic MCP Server independently.
* Adjust any budget at any time. Budgets are guardrails, not fixed entitlements.
* Notifications fire as daily consumption approaches and crosses a budget, so your administrators act before work stops.
* Usage rolls up against the yearly pack. There is no monthly reset, so a busy day balances against a quiet one.

### Where to see your credits

Your credit dashboard lives at [**secure.anvilogic.com/settings/agents/credit\_dashboard**](https://secure.anvilogic.com/settings/agents/credit_dashboard), under Settings, Agents, Credit Dashboard.

The dashboard shows:

* Credits consumed against your yearly pack, with a breakdown by product.
* Day-by-day consumption, so peak periods and credit-heavy workflows are easy to spot.
* Per-run records showing credits consumed and credits remaining after each run.
* Consumption by workflow for Blueprints, so owners see which automations carry the cost.

### Best practices

* Check the credit dashboard weekly. Consumption patterns shift as new Blueprints go into production.
* Set daily budgets on Blueprints and AI Chat before rolling either out broadly. Both are metered, and both are where consumption grows fastest.
* Size your pack against alert volume first. Threat Identifier and Threat Scenario verdicts are fixed rate, so that part of your annual consumption is straightforward to forecast.
* Revisit your budgets after any change to detection content or alert routing that moves alert volume.