AI

RAG multi-instance (Elasticsearch vector store)

Make documentation RAG (ai:index-docs / DocumentationService / DocumentationAgent) safe for multi-instance deployments by adding an Elasticsearch-backed VectorStoreInterface and clear deployment documentation, without changing the public CLI contract of ai:index-docs.

State
Done
Progress
100%
Steps
22 / 22
Started
2026-09-15
Finished
2026-09-15
Module
LaraAI

Tasks.

  • Task 1: Deployment documentation (filesystem interim)3/3
  • Task 2: Configuration keys for Elasticsearch RAG index3/3
  • Task 3: ElasticsearchRagVectorStore (core implementation)5/5
  • Task 4: Index template bootstrap (Artisan or documented curl)4/4
  • Task 5: Wire DocumentationAgent + isAvailable()5/5
  • Task 6: End-to-end checklist and README touch2/2

Read the full plan on GitHub