Skip to main content

Embed interface

The embedding interface converts text into vectors and is commonly used for semantic search, knowledge base retrieval, recommendation and deduplication.

Path​

POST /v1/embeddings

Base URL Please log in to the official website and view and copy it on the User Center - [API Key] page.

Usage suggestions​

  • Try to use the same embedding model in the same knowledge base.
  • Clean duplicate content and invalid text before storing it in the database.
  • Long text is first divided into paragraphs or semantic blocks.
  • Save the model name to facilitate subsequent index rebuilding.