Out of curiosity: how are embeddings handled in Pickaxe?!

If I understand correctly, every provider uses a different embedding model right so OpenAI’s embedding model would be different from Anthropic’s embedding model… but Pickaxe allows users to use any model they like, so is that to say you’re actually preparing embeddings of our documents for multiple embedding models at the same time? If so that is… impressive given how seamless the whole experience is.

We always use OpenAI’s embeddings model. And we cover the cost of it too.

1 Like