Skip to main content
POST
/
api
/
v1
/
runtime
/
vector-stores
Create vector store (CLI/SDK)
curl --request POST \
  --url https://api.runflow.ai/api/v1/runtime/vector-stores \
  --header 'Content-Type: application/json' \
  --data '
{
  "name": "<string>",
  "description": "<string>",
  "embeddingConfigId": "<string>",
  "llmProviderConfigId": "<string>",
  "llmModelConfigId": "<string>",
  "type": "KNOWLEDGE",
  "metadata": {},
  "chunkSize": 1000,
  "chunkOverlap": 200,
  "chunkStrategy": "recursive"
}
'

Documentation Index

Fetch the complete documentation index at: https://docs.runflow.ai/llms.txt

Use this file to discover all available pages before exploring further.

Body

application/json
​
name
string
required

Vector store name

​
description
string

Description

​
embeddingConfigId
string

Embedding configuration ID (optional if llmProviderConfigId + llmModelConfigId provided)

​
llmProviderConfigId
string

LLM Provider Config ID (for tenant-credential embeddings)

​
llmModelConfigId
string

LLM Model Config ID (for tenant-credential embeddings)

​
type
enum<string>
default:KNOWLEDGE

Vector store type

Available options:
KNOWLEDGE,
STRUCTURED
​
metadata
object

Custom metadata

​
structuredType
enum<string>

Structured document type (only for STRUCTURED stores)

Available options:
FAQ,
SURVEY,
GLOSSARY,
CATALOG,
QA,
DEFINITION
​
chunkSize
number
default:1000

Chunk size in tokens (default: 1000)

Required range: 100 <= x <= 8000
​
chunkOverlap
number
default:200

Chunk overlap in tokens (default: 200)

Required range: 0 <= x <= 1000
​
chunkStrategy
enum<string>
default:recursive

Chunking strategy (default: recursive)

Available options:
fixed,
recursive,
semantic

Response

201

Vector store created successfully.