Skip to main content
Pinecone is a vector database. MantrixFlow can write to Pinecone when your data includes an embedding column. Produce embeddings upstream (e.g. using a transformation or external service) before writing to this destination.

​
Prerequisites

You need a Pinecone account and an API key. Create an index with the appropriate dimensions for your embeddings.

​
Connection Setup

1

​
1. API Key

Obtain an API key from the Pinecone console. Paste it into the credentials field.
2

​
2. Index Name

Enter the name of the Pinecone index. The index must exist and have dimensions matching your embedding column.
3

​
3. Environment

Enter the Pinecone environment (e.g. us-east-1-aws). Find this in the Pinecone console.

​
Data Requirements

Your data must include:
  • Embedding column — A vector column (array of floats) with the same dimensions as the index.
  • Metadata columns — Optional. Stored as metadata in Pinecone.
  • Upsert key — Used as the vector ID for upsert operations. Must be unique per vector.