RAG: Search (vector)

rag_search · action · Utility · Available · v1.0.0

Description

Semantic retrieval over the TENANT vector knowledge base: embeds the query (OpenAI/Voyage/Ollama, BYOK) and runs cosine KNN on the selected vector DB, returning the top-k most similar chunks as grounding context for an agent/LLM node. Strict tenant isolation.

⚙️ Configuration parameters

Fields shown in the editor when configuring the node. Generated directly from the NodeDefconfigFields.

FieldTypeRequiredDefaultDescription
databaseId
Vector DB
db-pickeryesUn database con engine vettoriale (embedded o pgvector).
collection
Collezione
stringyesNome della collezione vettoriale dentro il DB.
query
Query
stringnoTesto da cercare. Vuoto = usa l'input del nodo precedente (stringa o {query}/{text}). Supporta {{espressioni}}.
provider
Provider embedding
enum
openaivoyageollama
noopenaiDEVE combaciare col provider/modello usato in ingest (stesso spazio vettoriale).
model
Modello embedding
stringnotext-embedding-3-smallStesso modello dell'ingest, altrimenti i vettori non sono confrontabili.
apiKey
API key (BYOK)
stringnoChiave del provider embedding (per openai/voyage). Ollama self-hosted non la richiede.
topK
Top K risultati
numberno5
minScore
Similarità minima (0-1)
numbernoScarta i risultati sotto questa soglia. Vuoto = nessuna soglia.
filterJson
Filtro payload
key-valuenoCoppie chiave-valore: ritorna solo i chunk col payload corrispondente.

💡 Configuration example

JSON snippet of the node as it appears in the workflow. Values are derived fromdefaultValue and from required parameters.

{
  "id": "node-rag_search-1",
  "defId": "rag_search",
  "label": "RAG: Search (vector)",
  "config": {
    "databaseId": "<databaseId>",
    "collection": "<collection>",
    "provider": "openai",
    "model": "text-embedding-3-small",
    "topK": 5
  }
}

🔗 Related nodes in the same category

Ready to use RAG: Search (vector)?

Available now on all FlowForge plans. Try it free without a credit card.

Start freeBrowse all nodes