# PartitionEmbeddingCredential

One of your own provider API keys that a partition's embeddings may authenticate with, in the partition's ordered list of keys. API keys only. Every key in the list must be for the provider of the partition's embedding profile.

**Kind:** an object, used in requests and responses. In a request it is called `PartitionEmbeddingCredentialInput`.

## In a request

| Field | Type | Required | Description |
|---|---|---|---|
| `provider` | [`EmbeddingProvider`](https://docs.maitento.com/reference/types/embedding-provider.md)`\| null` |  | The provider this key authenticates with. Null means `OpenAi`. |
| `apiKey` | `string` |  | The API key. Write-only: it is never returned. Write-only: never returned. |
| `baseUrl` | `string \| null` |  | The endpoint to call with this key, such as your own Azure OpenAI resource. Null or blank: the provider's default endpoint. Required for Qwen, which has none: give your region's workspace host as an absolute https `…/compatible-mode/v1` URL. The endpoint must serve the partition's model at the partition's width, or the embedding is rejected. |
| `modelIdentifier` | `string \| null` |  | The model or deployment id to send instead of the model's name, typically an Azure OpenAI deployment name. Null or blank: the model's name. |
| `credentialId` | `string` |  | Your own id for this key, so a failure recorded against it (`embeddingCredentialFailures`) can be matched back to the key you hold. |
| `priority` | `integer` |  | Failover order; lower is tried first. |

```json
{
  "provider": "OpenAi",
  "apiKey": "...",
  "baseUrl": "...",
  "modelIdentifier": "...",
  "credentialId": "...",
  "priority": 0
}
```

## In a response

Every field is present in a response; empty ones are `null`.

| Field | Type | Description |
|---|---|---|
| `provider` | [`EmbeddingProvider`](https://docs.maitento.com/reference/types/embedding-provider.md)`\| null` | The provider this key authenticates with. Null means `OpenAi`. |
| `baseUrl` | `string \| null` | The endpoint to call with this key, such as your own Azure OpenAI resource. Null or blank: the provider's default endpoint. Required for Qwen, which has none: give your region's workspace host as an absolute https `…/compatible-mode/v1` URL. The endpoint must serve the partition's model at the partition's width, or the embedding is rejected. |
| `modelIdentifier` | `string \| null` | The model or deployment id to send instead of the model's name, typically an Azure OpenAI deployment name. Null or blank: the model's name. |
| `credentialId` | `string` | Your own id for this key, so a failure recorded against it (`embeddingCredentialFailures`) can be matched back to the key you hold. |
| `priority` | `integer` | Failover order; lower is tried first. |

```json
{
  "provider": "OpenAi",
  "baseUrl": "...",
  "modelIdentifier": "...",
  "credentialId": "...",
  "priority": 0
}
```

## Where it's used

- Inside [`Partition`](https://docs.maitento.com/reference/types/partition.md)
- Inside [`PartitionEmbeddingCredentialsRequest`](https://docs.maitento.com/reference/types/partition-embedding-credentials-request.md)
- Inside [`PartitionEmbeddingProbeRequest`](https://docs.maitento.com/reference/types/partition-embedding-probe-request.md)
- Inside [`PartitionEmbeddingProfileRequest`](https://docs.maitento.com/reference/types/partition-embedding-profile-request.md)