orchardcore-ai-memory-elasticsearch

Orchard Core AI Memory Elasticsearch

Configure Elasticsearch memory indexing

You are an Orchard Core expert. Configure Elasticsearch as the persistent AI Memory vector backend while maintaining user isolation, stable provider mappings, and embedding compatibility.

Guidelines

• Enable the exact feature ID CrestApps.OrchardCore.AI.Memory.Elasticsearch.
• Its manifest depends on CrestApps.OrchardCore.AI.Memory and OrchardCore.Elasticsearch; the base memory feature is enabled by dependency.
• Create AI Memory (Elasticsearch) from Search → Indexing, select an embedding deployment, and choose it in Settings → Artificial Intelligence → Memory.
• The provider is registered as keyed IMemoryVectorSearchService using ElasticsearchConstants.ProviderName.
• Vector similarity never replaces authorization. Every retrieval must remain filtered to the current authenticated userId.
• Install this package in the web or startup project and secure the Elasticsearch connection with production secret configuration.
• Store only durable non-sensitive preferences and facts; do not persist credentials, tokens, financial data, or private keys as AI memory.

Provider registrations

Enable the backend

{
  "steps": [
    {
      "name": "Feature",
      "enable": [
        "CrestApps.OrchardCore.AI",
        "CrestApps.OrchardCore.AI.Memory.Elasticsearch",
        "OrchardCore.Elasticsearch"
      ],
      "disable": []
    }
  ]
}

The core AI Memory feature comes from the provider dependency. Use the Azure AI Search provider instead when that is the selected index service; do not configure both as masters unless the application intentionally manages separate indexes and migration.