Vector store resource

The vectorStore: resource adds documents to, and runs similarity search against, an external vector database. Unlike embedding: (SQLite keyword/vector index built into kdeps), vectorStore: talks to a real vector database service - Qdrant, Chroma, Pinecone, pgvector, and more - for production-scale RAG.

Where it runs

Both workflow mode and agent mode.

Basic usage

# resources/index.yaml
actionId: index
name: Index Documents
vectorStore:
  provider: qdrant
  url: http://localhost:6333
  collection: docs
  operation: add_documents
  documents:
    - content: "kdeps is a YAML-based AI agent framework."
      metadata: { source: readme }
  embedModel: text-embedding-3-small

# resources/search.yaml
actionId: search
name: Semantic Search
requires: [index]
vectorStore:
  provider: qdrant
  url: http://localhost:6333
  collection: docs
  operation: similarity_search
  query: "{{ get('q') }}"
  topK: 5
  embedModel: text-embedding-3-small

Supported providers

provider	Connection via url
qdrant (default)	http://localhost:6333
chroma	http://localhost:8000 (default if url is empty)
pinecone	https://<index-host>.svc.<env>.pinecone.io
azureaisearch	https://<service>.search.windows.net, or AZURE_AI_SEARCH_ENDPOINT env var
opensearch / elasticsearch	http://localhost:9200
weaviate	http://localhost:8080
pgvector / postgres / postgresql / alloydb / cloudsql	PostgreSQL DSN, e.g. postgres://user:pass@localhost/db
mariadb / dolt / mysql	MySQL DSN, e.g. user:pass@tcp(localhost:3306)/dbname
mongodb / mongo	MongoDB URI, e.g. mongodb://localhost:27017
redis	Redis URI, e.g. redis://localhost:6379 (default if url is empty). collection is the Redis index name
bedrock	AWS Bedrock Knowledge Base - collection is the knowledge base ID; no url or embedModel needed, embedding happens server-side. Uses the standard AWS SDK credential chain (AWS_ACCESS_KEY_ID, AWS_SECRET_ACCESS_KEY, AWS_REGION)

Configuration options

Option	Operation	Description
provider	all	Vector store backend (see table above)
url	all	Endpoint or DSN for the store
collection	all	Collection / index / table name (required)
apiKey	all	Auth token. For mongodb/mongo, used as the database name instead (default "kdeps"). For opensearch/elasticsearch, format "user:pass"
operation	all	add_documents or similarity_search (required)
documents	add_documents	List of { content, metadata } documents to upsert
query	similarity_search	Natural language search query
topK	similarity_search	Number of results to return (default: 5)
embedModel	all	Embedding model used to vectorize documents/queries, e.g. text-embedding-3-small (required)
embedBackend	all	Embedding provider: openai, ollama, google
embedBaseURL	all	Custom base URL for an OpenAI-compatible embedding backend

Output

add_documents:

{ "success": true, "upserted": 1 }

similarity_search:

{
  "success": true,
  "results": [
    { "content": "kdeps is a YAML-based AI agent framework.", "metadata": { "source": "readme" }, "score": 0.91 }
  ]
}

Feeding from a loader

loader:'s output documents map directly onto vectorStore:'s documents: field - see the Loader Resource for the full load-then-index pipeline.

See also

• Loader resource - load files/URLs into documents before indexing
• Embedding resource - local SQLite-backed alternative, no external service required
• Resources overview - all resource types