Embeddings are now supported

BrushLLM now serves embeddings alongside chat models. Text Embedding v4 is a multilingual embedding model that fits search and RAG pipelines, callable through the same base URL and API key as every other model.

The request shape follows the OpenAI embeddings API exactly: POST /v1/embeddings with your model id and input. Check the Models page for the current price per 1M tokens.

This unlocks the other half of the AI stack — retrieval — without leaving the gateway. More embedding sizes and languages are on the roadmap.