> For the complete documentation index, see [llms.txt](https://anyint.gitbook.io/docs/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://anyint.gitbook.io/docs/pricing-and-plans/custom-subscriptions.md).

# custom subscriptions

Custom subscriptions let you purchase access model by model instead of buying one shared bundle. Each selected model has its own limits and service-level settings.

## What you configure per model

* model ID
* service tier
* daily call limit
* monthly usage cap

For text models, monthly caps are usually token-based. For image or video models, monthly caps are typically generation-count based.

## Service tiers

The current subscription design exposes three user-facing levels:

| Tier         | Positioning             |
| ------------ | ----------------------- |
| `basic`      | lowest-cost option      |
| `stable`     | balanced option         |
| `enterprise` | highest-priority option |

Explain these tiers to customers as service-level choices, not as internal upstream key-pool mechanics.

## Why teams choose custom subscriptions

* only pay for the models they actually use
* give different workloads different ceilings
* separate text, image, video, and premium-model usage
* align spend with specific teams or products

## Important rules

* one account should not hold an active plan subscription and an active custom subscription at the same time
* a user should have at most one active custom subscription
* model access is restricted to what the custom subscription explicitly includes
* renewing an expired custom subscription should issue a new key and reset usage from zero

## Example purchase shape

```json
{
  "models": [
    {
      "modelId": "claude-opus-4-6",
      "channelTier": "stable",
      "dailyCallLimit": 50,
      "monthlyTokenCap": 5000000
    },
    {
      "modelId": "nano-banana-2",
      "channelTier": "basic",
      "dailyCallLimit": 20,
      "monthlyTokenCap": 200
    }
  ]
}
```

## Customer-facing expectation

When a customer buys a specific service tier for a specific model, AnyInt should respect that contract. If that tier is unavailable, requests may fail rather than silently move to a different tier with different commercial or reliability characteristics.