AI Agents & MCPs & AI Workflow Automation • (~400 MCP servers for AI agents) • AI Automation / AI Agent with MCPs • AI Workflows & AI Agents • MCPs for AI Agents
TypeScript ai-agentai-agent-toolsai-agentsai-agents-frameworkmcpmcp-servermcp-toolsmcpsn8n-alternativeno-code-automationworkflowworkflow-automationworkflows
---
title: "AI usage"
description: "See where your AI spend goes across projects, flows, agents, chat, keys and models"
icon: "chart-pie"
---
## Overview
**AI usage** answers one question: where is your AI money going? It covers every AI call on the platform, from flow steps, agents, chat and knowledge bases, on your own keys and on Activepieces credits.
Open it at `/platform/ai/usage`.

<Note>
AI usage is available on Cloud and self-hosted Enterprise. It is not in the menu yet, so open it by URL. Community Edition doesn't have it. On self-hosted Enterprise, the page needs a billing connection; without one it says AI usage isn't available.
</Note>
New usage shows up a few minutes after the AI call.
## Period and scope
Two controls sit next to the page title, on every usage page:
- **Scope**: **Production** (the default) counts live flows, agents and chat. **Testing** counts test runs from the flow builder and the agent builder. **All** counts both.
- **Period**: the last hour, 6 hours, 12 hours, 24 hours, 7 days, 30 days or 90 days, this month or last month.
The whole page follows both. They stay in the URL, so you can share a view.
## Overview page
From top to bottom:
1. **Tiles**: the pace of spend, AI calls and tokens.
- For a period of a day or less, the pace is a **Daily run rate**: the spend of the period, scaled to a full day.
- For longer periods it is the **Average per day**.
- For **This month** it is a **Month-end forecast**, with the daily pace it is based on.
2. **Chart**: total spend for the period, the change from the previous period, and a line chart. The solid line is this period, the dashed line the previous one. Switch between **Spend**, **Calls** and **Tokens**.
3. **Spend by source**: what kind of work spends: **Flow AI steps**, **Run Agent steps**, **Agents**, **Chat** and **Knowledge base**.
4. **Top lists**: the top five **Projects**, **Flows**, **Agents**, **Users**, **Providers & keys** and **Models**. The bar under each row is its share of all AI spend in the period.
"Previous period" is the same length of time right before. **This month** compares with the same days of last month, and **Last month** with the month before it. If tracking started after the previous period began, the change is hidden, so a new platform never shows fake growth.
## Lists and details
Click **See all** on a card to see the full list. Search by name, click a column header to sort, and page through the results.
Click any row to open its detail page. It shows the same chart, the top five of the other lists (for example, a flow's models and keys), and one extra card for its type:
| Detail page | Extra card |
| --- | --- |
| Project | Spend, with its number of flows and agents |
| Flow | Runs and cost per run |
| Key | Cost reported by the provider vs estimated, and a link to the provider's dashboard |
| Model | List price per million tokens |
The page title is a breadcrumb, for example **AI usage / Flows / Lead enrichment**. Click a part of it to go back.
A Run Agent step counts under both its flow and its agent, so flows and agents together can add up to more than a project's total.
## Estimated vs exact cost
All amounts are in dollars. Activepieces credits are converted to their dollar value.
Click **Estimated** next to the spend to see why:
- Most providers return how many tokens a call used, not what it cost. Activepieces multiplies the tokens by the model's list price.
- Discounts, cached tokens, batch rates and price changes on your account can make your real bill a few percent different.
- When a provider does return the exact cost, it is used as is. A key's detail page shows how much was reported and how much was estimated.
- Activepieces credits are always exact.
For billing, your provider's own dashboard is the source of truth.
## Related
- [Providers and keys](/admin-guide/guides/setup-ai-providers)
- [Tiers](/admin-guide/guides/ai-tiers)
- [Usage](/admin-guide/guides/usage) for runs and credits on your plan.
- API reference: [overview](/endpoints/ai-usage/overview), [list items](/endpoints/ai-usage/list-items), [get an item](/endpoints/ai-usage/get-item).