Aliyun OSS Vector Store icon

Aliyun OSS Vector Store

Store and retrieve vectors using Alibaba Cloud OSS Vector Bucket

Actions7

Retrieve (as Tool for AI Agent)

AI-generated

Summary

Configure the node as a tool for an AI Agent to perform vector similarity searches against an Alibaba Cloud OSS vector index. It uses a connected embeddings model to embed the query extracted from the AI Agent's tool call arguments, searches the OSS vector index, and returns the most relevant text chunks formatted as tool response items.

Inputs

  • Index Name (required) — The OSS vector index name to query, which must correspond to an existing index in your OSS bucket.
  • Tool Name (required) — The alphanumeric name (underscores allowed) by which the AI Agent identifies and calls this tool.
  • Tool Description (required) — A specific text description explaining the purpose of this vector search tool to the AI Agent.
  • Limit — Maximum number of vector search results to return (default 4).
  • Include Metadata — Whether to include document metadata in each returned search result; page content is always included (default true).
  • Metadata Filter — Optional JSON filter using OSS MongoDB-style operators to restrict search results by metadata fields.

Output shape

a list of tool response objects, each containing an array of text chunks (JSON strings) with page content and optionally metadata

The node returns one output item per input item containing the search results. If no query can be extracted from the AI Agent's tool call arguments, an error message is returned in the response array advising no further calls with empty arguments. If no results are found, a special message signals the AI Agent to stop further retrieval attempts and answer based on general knowledge.

Links

Discussion