embedding-upload

To generate server-side embeddings from a CSV file, you upload the file to process your data in batches, producing 3072-dimensional embeddings per window alongside metadata like query_timestamp and read_index.

Yes, you can extract CSV embeddings using both Python and web clients. The server-side processing supports both environments and can stream results via SSE for real-time data delivery.

The dimension size of the embeddings produced from CSV data is 3072-dimensional per window, generated from 4×768 inputs. This output includes associated metadata for downstream analysis.

Yes, you need an API key to upload CSVs for embedding extraction. The lens-based workflow requires this key to authenticate requests and process your data securely on the server side.

When generating embeddings from a CSV upload, the metadata included provides query_timestamp and read_index. This associated metadata is returned with the 3072-dimensional embeddings for downstream analysis.

Processing CSV data for server-side embeddings processes inputs in 4×768 windows to produce 3072-dimensional outputs. You must manage API key authentication and handle streaming via SSE for large files.