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

# Plans & Limits

> Compare Declaw Cloud tiers. Every limit below is enforced at the API gateway — requests that exceed them are rejected with a documented error code.

Declaw Cloud has three tiers. Free is self-serve with no spend commitment; Pro is self-serve with a monthly recharge minimum; Enterprise is contact-sales.

## Comparison

|                                      | Free      | Pro          | Enterprise    |
| ------------------------------------ | --------- | ------------ | ------------- |
| **Concurrent sandboxes**             | 25        | 500          | Custom        |
| **Max session duration**             | 1 hour    | 72 hours     | Custom        |
| **Max vCPUs / sandbox**              | 4         | 16           | Custom        |
| **Max memory / sandbox**             | 4 GB      | 16 GB        | Custom        |
| **Max disk / sandbox**               | 10 GB     | 50 GB        | Custom        |
| **Max egress connections / sandbox** | 50        | 200          | Custom        |
| **General requests / sec**           | 500       | 2,000        | Custom        |
| **Sandbox creates / sec**            | 2         | 10           | Custom        |
| **Deposit range**                    | $5 – $100 | $10 – $5,000 | Custom        |
| **Minimum monthly recharge**         | —         | \$100        | Custom        |
| **Self-serve upgrade**               | ✓         | ✓            | Contact sales |

<Note>
  Enterprise is not self-selectable from your dashboard — [contact sales](mailto:team@declaw.ai) to upgrade.
</Note>

## How limits are enforced

Every limit above maps to a specific point in the request pipeline. When you hit one, the API returns a documented error so your client can react:

| Limit hit                                       | HTTP status |
| ----------------------------------------------- | ----------- |
| Concurrent sandboxes                            | `429`       |
| Session duration (request `timeout` > tier max) | `403`       |
| vCPUs / memory / disk (template or request)     | `403`       |
| General request rate                            | `429`       |
| Sandbox create rate                             | `429`       |

Response bodies are always JSON of shape `{"message": "<reason>"}`. The specific `<reason>` strings you'll see:

```json theme={null}
{ "message": "concurrent sandbox limit reached for tier: pro" }
{ "message": "session duration exceeds tier limit" }
{ "message": "vcpu limit exceeds tier allowance" }
{ "message": "memory limit exceeds tier allowance" }
{ "message": "disk limit exceeds tier allowance" }
{ "message": "rate limit exceeded" }
{ "message": "sandbox creation rate limit exceeded" }
```

See [Errors & Rate Limits](/platform/errors) for full handling guidance.

## Recharge commitment

**Pro** requires a minimum *total deposit* of \$100 per calendar month. **Enterprise** commitments are negotiated per contract. This is a spend commitment, not a flat fee — any deposits you already made in the month count toward it. You pay only for what you consume; the unused balance carries over.

If you miss a monthly commitment, the account enters a **grace period** (duration set per account by billing). During grace, your tier limits remain active. After grace expires without a qualifying deposit, the account is downgraded to Free.

## Free tier sandbox balance

Free accounts do not require a positive paid balance to create sandboxes — they consume the **free credit grants** described in [Billing & Pricing](/platform/billing). Once the free grant is exhausted, creation requests return `402 Payment Required` until you deposit funds or upgrade.

Pro and Enterprise require a positive wallet balance (free credits + paid balance combined) for any billable operation.