---
title: 'Pricing'
description: 'What Flexkit charges for: plans, included usage, on-demand rates, AI credits and taxes.'
---

import { Callout } from 'nextra/components';

# Pricing

Flexkit bills each **team**, for the projects it owns. Every project is on its own [plan](/docs/plans): **Free** or **Pro**. A team can mix both, and each project's usage counts only against its own allowance.

There are three kinds of charges:

| Charge              | Applies to                                | Billed                                                                 |
| ------------------- | ----------------------------------------- | ---------------------------------------------------------------------- |
| **Plan seats**      | Pro projects                              | Monthly, in advance                                                    |
| **On-demand usage** | Pro projects, beyond the included usage   | Monthly, in arrears                                                    |
| **AI credits**      | Any project's AI model usage (LLM tokens) | Prepaid, from the team's [AI credit](/docs/billing/ai-credits) balance |

All prices are in US dollars and exclude taxes.

## Plans

| Plan       | Price                                                                         |
| ---------- | ----------------------------------------------------------------------------- |
| Free       | $0                                                                            |
| Pro        | $25 per seat per month, per project                                           |
| Enterprise | Coming soon: custom pricing. See [plans](/docs/plans#enterprise-coming-soon). |

A Pro project's **seats** are its members, except Viewers, with a minimum of one. Adding a member during the month charges the rest of that month right away (prorated). Removing one lowers the next invoice.

## Included usage

Each project gets this usage every calendar month. Usage resets on the 1st, at 00:00 UTC.

| Resource          |  Free   |   Pro   |
| ----------------- | :-----: | :-----: |
| API requests      | 250,000 | 250,000 |
| Database capacity | 0.5 GB  | 0.5 GB  |
| Asset storage     |  50 GB  |  50 GB  |
| Agent chat turns  |   50    |   300   |
| Automation runs   |   10    |   100   |
| Agent steps       |   500   |  3,000  |

The two plans differ in what happens **beyond** the included usage:

- **Free:** at 100% the project becomes read-only, or the agent stops. Nothing is charged. See [plans](/docs/plans#what-happens-when-i-reach-100).
- **Pro:** usage continues and is charged at the on-demand rates below. [Spend management](/docs/spend-management) lets you set a monthly limit, or turn on-demand usage off.

## On-demand rates

| Resource          | Rate                           |
| ----------------- | ------------------------------ |
| API requests      | $4 per 100,000 requests        |
| Database capacity | $25 per GB-month               |
| Asset storage     | $0.50 per GB-month             |
| Agent chat turns  | $10 per 1,000 ($0.01 per turn) |
| Automation runs   | $20 per 1,000 ($0.02 per run)  |
| Agent steps       | $5 per 1,000 ($0.005 per step) |
| Sandbox compute   | At cost                        |

On-demand usage appears on the invoice issued at the end of each month, next to the following month's seats.

## How usage is calculated

**API requests** count every request made to the project's APIs during the month.

**Database capacity** and **asset storage** are billed as a **time-weighted average** (GB-months). Flexkit measures them every hour and weighs each hour by its share of the month. You pay for how much you stored, and for how long, not for a peak.

For example, a Pro project 1 GB above its 0.5 GB database capacity for half the month uses 0.5 GB-month of on-demand capacity: 0.5 × $25 = **$12.50**.

**Agent chat turns** count each message the agent answers in [agent chat](/docs/ai/agent-chat). **Automation runs** count each run of an [automation](/docs/automations). **Agent steps** count each model call the agent makes, together with the tools it runs. A short answer is one step; a long investigation can take dozens.

**Sandbox compute** is the code-execution sandbox the agent uses for some tasks. It's billed at what the sandbox provider charges Flexkit: active CPU time, memory, and sandbox creation.

## AI credits

Model usage (LLM tokens) in agent chat, automations, Studio AI, editor completions and voice transcription is paid from the team's **AI credit** balance, not from the plan or on-demand usage.

Each model call costs exactly what the AI provider charges for it, with **no markup**. Credits are prepaid, and the balance can top itself up automatically. New users get **$5 of free credit** once, on their first team. See [AI credits](/docs/billing/ai-credits).

## Taxes

Prices exclude taxes. Where Flexkit has to collect sales tax or VAT, it's added to invoices and credit purchases based on the team's billing address and tax ID. Set both in [billing settings](/docs/billing#billing-details).

<Callout type="info">
  Charges are shown in the dashboard as they accrue. The project's **Usage** page and **Settings → Spending** show this
  month's on-demand usage so far. See [usage](/docs/usage).
</Callout>

## More resources

- [Plans](/docs/plans): what Free and Pro include, and how to change plans.
- [Spend management](/docs/spend-management): limit or turn off on-demand usage.
- [Limits](/docs/limits): every limit in one place.
- [Billing](/docs/billing): payment methods, invoices, and failed payments.


---

[View full sitemap](/docs/sitemap.md)