Skip to main content

TUTORIAL

Odin Tutorials

From workspaces to agents, knowledge bases and automation: learn Odin step by step.

Embedding Model

Add and manage the embedding models used for semantic search and retrieval.

Add and manage the embedding models used for semantic search and retrieval. There's a search box at the top of the page, and a New Embedding Model button in the top right.

Embedding Model list, one built-in Balanced-tier model

  1. Built-in models​

    Unlike Completion Model, Embedding Model currently ships with only one tiered model, marked Built-in and not deletable or editable:

    • Builtin (Balanced) — general-purpose embedding model

  2. Adding a custom model​

    Clicking the New Embedding Model button in the top right opens a form. Unlike the Completion Model form, the fields here change depending on Model Provider: the default selection is Azure OpenAI Embedding Model, which requires Name (required), Model Provider (dropdown), Resource Name (required), Deployment ID (required), API Version (required), and API Key (required) — six fields in total, more than Completion Model's form. Click Save to save, or Cancel to discard.

    New Embedding Model form, default provider is Azure OpenAI Embedding Model, all six fields required

Ready to get started?

Want to see how Asgard fits your team? Talk to us.

Contact us