Qdrant Store AI Tools icon

Qdrant Store AI Tools

Embed text and binary files into a Qdrant vector store using a connected embedding model. Binary files are sent directly to the embedder (multimodal support with automatic fallback). Connect an Embedding model to the "Embedding" input and the "Tool" output to an AI Agent.

Qdrant Store AI Tools

AI-generated

Summary

Provide a set of AI tools to embed and store text strings, binary files, or multiple texts in batches as vector points within a specified Qdrant collection, as well as manage Qdrant collections by creating or deleting them.

Inputs

  • defaultCollection (required) — Default Qdrant collection name to store vectors in. Must exist or be created via the "Create Collection" tool.
  • enabledTools (required) — Select which AI tools to expose to the AI Agent. Options include: storing individual text, storing a binary file, batch storing texts, creating a collection, and deleting a collection.

Output shape

a list of AI tools respecting the enabledTools setting, each callable with JSON parameters to perform vector store operations or collection management

Each tool returns a JSON stringified object indicating success or error and relevant details such as stored point IDs or collection info. Tools require a connected embedding model sub-node at the "Embedding" input. Binary embedding attempts multiple fallback strategies for compatibility. The "Tool" output is meant for connection to an AI Agent node's "Tools" input.

Examples

Example 1: Embedding and storing a single text string as a vector point

Enable the 'Store Text' tool and specify the defaultCollection name. Call the 'qdrant_store_text' tool with parameters 'text' (string to embed), optionally 'collection' to override default, optional 'point_id', and optional 'metadata'. The text is embedded using the connected embedding model and stored in Qdrant with an auto-generated UUID if no point ID is supplied. Returns the stored point's ID and vector size.

Example 2: Embedding and storing a binary file

Enable the 'Store Binary File' tool and specify the defaultCollection. Provide a binary property name referencing the file in workflow binary data. Optionally specify 'collection', 'point_id', 'chunk_size' (for splitting large text files), and 'metadata'. The node tries to embed the raw binary directly; if embedding fails, falls back to base64 data URI string or descriptive text embedding. Binary data is stored as vector points in Qdrant, chunked if applicable. Returns stored point IDs and count.

Example 3: Batch embedding and storing multiple texts

Enable the 'Batch Store Texts' tool. Pass an array of text strings as 'texts', optional 'collection', optional array 'metadatas' aligned with texts, and optional 'id_prefix'. All texts are embedded in one call and upserted in Qdrant at once. Returns count of stored points and their IDs.

Links

Discussion