# Exoscale Documentation — Service Boundaries - Concrete AI

> Concrete AI limits, quotas, pricing, and links to product-specific service boundaries.

---

url: https://community.exoscale.com/product/concrete-ai/service-boundaries/

---

# Service Boundaries - Concrete AI

- [Limits and Quotas - AI API Keys](https://community.exoscale.com/product/concrete-ai/service-boundaries/limits-and-quotas/index.md) - AI API Key limits and quotas.
- [SLA - Concrete AI](https://community.exoscale.com/product/concrete-ai/service-boundaries/sla/index.md) - Concrete AI Service Level Agreement

- [Pricing ↗︎](https://www.exoscale.com/pricing/#inference)

---

url: https://community.exoscale.com/product/concrete-ai/service-boundaries/limits-and-quotas/

---

# Limits and Quotas - AI API Keys

## Limits

Name format
: Key names are 1 to 50 characters. Letters, digits, spaces and `_'()-` are accepted, but the name must start and end with a letter or digit.

Scope rules
: `models` and `deployments` each accept explicit entries. Deployments in `error` or `deleted` state cannot be added; any other state (including deployments still being provisioned) is accepted. Models must be public models.

Plaintext shown once
: The key value is returned only in the `create-ai-api-key` response. If you lose it, create a new key and revoke the old one.

Revocation retention
: A revoked key stays listed with `revoked-at` set for 30 days, then is deleted. Revoked keys cannot be updated.

Zone scope
: Keys are stored per zone. A key is created in the same zone where the deployment or on-demand inference end-point is hosted. If it is, for example, `de-fra-1` , it authenticates endpoints in `de-fra-1` only.

## Quotas

Key count
: The number of AI API Keys per organization is limited by quota. `create-ai-api-key` consumes one unit and `revoke-ai-api-key` releases it. To request an increase, please contact support by opening a Ticket in the Portal.

---

url: https://community.exoscale.com/product/concrete-ai/service-boundaries/sla/

---

# SLA - Concrete AI

## Concrete AI Service-Level Agreement (SLA)

Concrete AI combines services with separate Service Availability Targets. This page brings together the applicable commitments; the product-specific SLA pages below contain the complete terms, exclusions, client responsibilities, and service-credit conditions.

### Availability Commitments

| Service | Target | Scope |
| --------- | -------- | ------- |
| Dedicated Inference | >= 99.95% Monthly Uptime Percentage | Dedicated Inference endpoint availability, calculated per Client and endpoint. |
| GPU Servers | >= 99.95% Monthly Uptime Percentage | Availability of each Exoscale GPU Server instance. |
| Managed vector databases | Varies by DBaaS plan | Managed PostgreSQL and OpenSearch vector capabilities follow the applicable DBaaS plan SLA. |

## Product-Specific SLAs

- [Dedicated Inference SLA](https://community.exoscale.com/product/concrete-ai/dedicated-inference/service-boundaries/sla/index.md)
- [GPU Servers SLA](https://community.exoscale.com/product/compute/instances/service-boundaries/sla/index.md)
- [DBaaS SLA](https://community.exoscale.com/product/dbaas/service-boundaries/sla/index.md)