Aliyun OSS Vector Store icon

Aliyun OSS Vector Store

Store and retrieve vectors using Alibaba Cloud OSS Vector Bucket

Actions7

Insert

AI-generated

Summary

Insert documents by embedding their content into vector representations and storing these vectors in the specified Alibaba Cloud OSS vector index.

Inputs

  • Index Name (required) — The name of the OSS vector index to write to; must match an existing index in your bucket or will be created if 'Auto Create Index' is enabled.
  • Auto Create Index — Whether to automatically create the vector index if it does not already exist.
  • Vector Dimension — The embedding vector dimension (e.g., 2560); must match the embedding model output and the index's configuration. Required if 'Auto Create Index' is true.

Output shape

A list of output items corresponding to each main input item, each containing metadata about the insertion operation such as success status, number of vectors inserted, and any failed item keys.

The node requires two additional inputs connected: an AI Embedding node to provide embeddings and an AI Document loader node that processes each item into chunks for embedding. Each chunk's text and metadata are embedded and stored as individual vectors. The node handles automatic batching and retries for ingestion. If an error occurs on a particular item, it logs the failure and continues processing remaining items. The output reports insertion success per input item. If no documents are extracted, zero vectors are inserted for that item.

Examples

Example 1: You have a collection of documents you want to add to an Aliyun OSS vector index and have an embedding model node and document loader node connected. Configure the node with the OSS index name, enable 'Auto Create Index' if needed, and specify the vector dimension matching your embedding model. The node will embed and upload document chunks as vectors to the OSS index, returning per-item success and inserted vector counts.

Set 'Index Name' to your target index, 'Auto Create Index' to true if the index may not exist, and 'Vector Dimension' to your embedding output size. Connect an Embeddings node and a compliant Document loader node as inputs, and trigger the node to insert the documents.

Links

Discussion