{
  "$schema": "https://json-schema.org/draft/2019-09/schema",
  "$id": "https://www.krakend.io/schema/v3.0/ai/semantic-cache.json",
  "title": "AI Semantic Cache",
  "description": "Configuration for the AI Semantic Cache, including the embedding models used to generate vector representations of requests.\n\nSee: https://www.krakend.io/docs/enterprise/ai-gateway/semantic-cache/",
  "type": "object",
  "properties": {
    "redis_connection": {
      "description": "The name of the [Redis connection](https://www.krakend.io/docs/enterprise/service-settings/redis-connection-pools/) to use, it must exist under the `redis` namespace at the service level and written exactly as declared.",
      "default": "default",
      "type": "string"
    },
    "embedder": {
      "description": "The name of the embedder to use, as defined in service extra_config",
      "default": "default",
      "type": "string"
    },
    "max_distance": {
      "description": "The maximum cosine distance allowed when searching for semantically similar vectors. A value of 0 represents identical vectors; larger values allow greater differences.",
      "default": 0.25,
      "type": "number"
    },
    "key_variants": {
      "description": "Request fields used to create distinct cache entries. Requests with different values for these fields are stored and matched separately, even when their vectorized content is identical.",
      "type": "array",
      "items": {
        "description": "A request field whose value is used to distinguish cache entries.",
        "examples": [
          "req_params.Ids",
          "req_headers.x-user",
          "req_query_string.user"
        ],
        "type": "string"
      }
    },
    "req_content": {
      "description": "The part of the request to vectorize. Use this to focus semantic matching on the content that can vary between requests. For example, if requests contain a large system prompt shared across calls, you can select only the user prompt to focus vector search on the variable content. If not specified, the entire request body is vectorized.",
      "examples": [
        "req_params.Ids",
        "req_headers.x-user",
        "req_query_string.user",
        "req_body.content.prompt"
      ],
      "type": "string"
    },
    "ttl": {
      "title": "Cache TTL",
      "description": "The time-to-live (TTL) for cached entries.",
      "default": "600s",
      "$ref": "../timeunits.json#/$defs/timeunit",
      "type": "string"
    }
  },
  "patternProperties": {
    "^[@$_#]": true
  },
  "additionalProperties": false
}