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

# Included AI

> AI comes with your plan: which models do the work, how your allowance and overage work, and how your content is handled.

AI is included with your plan. You don't connect, configure or pay for an AI provider to use Iron Octopus: from your first idea, Iron Octopus plans, builds and reviews with leading models from OpenAI and Anthropic. If you'd rather use your own AI subscription, you can. See [use your own AI provider](https://docs.ironoctopus.ai/ai/your-own-provider.md).

## What's included

* **The models.** Iron Octopus chooses the right model from OpenAI or Anthropic for each job: planning, building, reviewing and answering questions in Ask.
* **An AI allowance.** Your plan includes an allowance of AI usage, shown in your plan details when you subscribe. Usage is measured in tokens, the units AI providers charge by, and converted to cost at the provider rates listed in your plan terms.
* **Every AI worker.** The platform's AI workers use included AI by default. See [AI workers and teams](https://docs.ironoctopus.ai/ai/workers-and-teams.md).

## How usage is counted

Each AI request reserves an amount from your allowance before it runs, and settles at the actual cost when the provider reports it. That's why **Some provider measurements are pending** can appear on the Usage page for a short time.

The rates are listed in USD per million tokens, for each provider and model, under **Provider rates included in these terms**:

| Rate | What it's for |
| - | - |
| Input tokens, uncached | What Iron Octopus sends to the model. |
| Cache read | Input the model has seen recently, which costs less. |
| Cache write | Storing input for reuse. |
| Output | What the model writes back. |

## When the allowance runs out

When the included allowance is used up, Iron Octopus's AI work pauses until the next period's allowance or until overage is authorized. Your plans, code and records aren't affected.

To keep working past the allowance, the workspace **Owner** can authorize overage on the Usage page:

**1. Open Usage**

Go to Workspace settings › **Usage**, and find **Platform AI usage**.

**2. Authorize overage**

Click **Authorize overage**, accept the current terms, and enter a **monthly limit** in USD. AI work continues past the allowance up to that limit.

Each period's overage charges appear under **AI usage statements**. To stop, click **Revoke overage consent**. If the page says **Overage enrollment is not currently available**, overage can't be authorized yet, and work pauses at the end of the allowance.

## How your content is handled

* Iron Octopus sends the parts of your content each task needs to the AI provider, through the provider's business API.
* **No training.** Your content isn't used to train or fine-tune any AI model, and the providers' terms don't allow them to train on it.
* Providers may keep requests briefly to detect abuse, then delete them.

See [security](https://docs.ironoctopus.ai/security/overview.md) for more.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.