Actions10
- Orchestration Actions
- Vector Store Actions
Vector Store → Upsert Documents
AI-generatedSummary
Upsert documents into a vector store backend.
Inputs
- documentsJson (required) — Array of documents: [{
- vectorBackend — Choose where vectors are persisted between executions
- vectorDimension — Deterministic embedding dimension used for upsert and search
- cacheNamespace — Namespace key for local persistent cache
- qdrantUrl — Qdrant server URL
- qdrantCollection — Qdrant collection name
- qdrantApiKey — Optional for local/self-hosted Qdrant
- pineconeHost — Example: your-index-xxxx.svc.region.pinecone.io
- pineconeApiKey — Pinecone API key
- pineconeNamespace — Pinecone namespace
Output shape
An array of paired execution results, each containing a json payload and pairedItem index.
The returned
jsonobject varies byvectorBackendsetting. All outputs includeresource,operation, andbackendfields.
Examples
Example 1: Local cache backend
vectorBackend=localCache, documentsJson=[{content:'Hello',metadata:{source:'demo'}}] with default vectorDimension=64 and cacheNamespace='myNamespace' ;nReturns: {json:{resource:'Vector Store',operation:'Upsert Documents',backend:'localCache',namespace:'myNamespace',upsertedCount:1,storedCount:1},pairedItem:{item:0}}
Example 2: Qdrant backend
vectorBackend=qdrant, documentsJson=[{content:'Hello',metadata:{source:'demo'}}] with default qdrantUrl='http://localhost:6333', qdrantCollection='mycollection' ;nReturns: {json:{resource:'Vector Store',operation:'Upsert Documents',backend:'qdrant',collection:'mycollection',upsertedCount:1,status:'ok',operationId:null},pairedItem:{item:0}}
Example 3: Pinecone backend
vectorBackend=pinecone, documentsJson=[{content:'Hello',metadata:{source:'demo'}}] with pineconeHost='my-index-xyz.svc.us-west1.gcp.cloud.pinecone.io', pineconeNamespace='myNamespace' ;nReturns: {json:{resource:'Vector Store',operation:'Upsert Documents',backend:'pinecone',namespace:'myNamespace',upsertedCount:1},pairedItem:{item:0}}