AiSearchConfiguration
A request or response body; the type column links to the schemas it refers to.
AiSearchConfiguration
| Field | Type | Required | Description |
|---|---|---|---|
| enabled | boolean | yes | Whether `AI_SEARCH_EMBEDDING_PROVIDER` names a provider. |
| embedding_provider | string | null | yes | one of "openai" | "voyage" | "gemini" | "local" | null |
| embedding_model | string | null | yes | — |
| embedding_dimensions | integer | null | yes | The vector length, once known; discovered from the first embedding. |
| embedding_base_url | string | null | yes | The OpenAI-compatible endpoint in use, when not the provider's own. |
| embeds_images_natively | boolean | yes | Whether the embedding model takes images itself (multimodal). |
| media_provider | string | yes | one of "none" | "openai" | "gemini" | "anthropic" |
| media_model | string | null | yes | — |
| max_file_bytes | integer | yes | Files larger than this are indexed by their name alone. |
| problems | array of string | yes | Configuration the server could not use, as it would be logged. |
jsonExample
{
"enabled": true,
"embedding_provider": "openai",
"embedding_model": "text-embedding-3-small",
"embedding_dimensions": 0,
"embedding_base_url": "string",
"embeds_images_natively": true,
"media_provider": "none",
"media_model": "string",
"max_file_bytes": 0,
"problems": [
"string"
]
}